Anthropic pulled powerful AI models offline after U.S. export control order over security fears

"Dangerous" AI models are coming no matter what

Late last week, Anthropic took its new Claude Fable 5 and Mythos 5 AI models offline following a United States government export-control directive barring "any foreign national" from using the services. The company has been in talks with the White House since Friday but has yet to secure an agreement that would allow it to reinstate the offerings. Since Mythos debuted in April, Anthropic has claimed -- and warned -- that the model has advanced capabilities for not only finding software vulnerabilities to help defenders patch them, but also figuring out ways to exploit them that could be used by bad actors. Anthropic itself noted this double-edged sword in its launch of Mythos 5 and Claude Fable 5. "A great deal of advanced usage of AI models is dual use: the same queries that are beneficial in the hands of cybersecurity professionals and biology researchers could be dangerous if available to malicious actors," the company wrote in a blog post last week. With this in mind, the company initially released a version called Mythos Preview to a select consortium as part of a working group known as Project Glasswing. Mythos 5 was also privately released to this group last week, while Claude Fable 5, which is a Mythos-grade model, was released to the general public with specific blocks on its ability to give responses to questions about biology and cybersecurity. Then, at the end of last week, the Trump administration moved to restrict both models because it believes that Fable 5's guardrails can be disabled to allow full access to the Mythos 5 capabilities, allegedly making it a national security risk. Experts say, though, that this institutional clash is simply delaying or masking a hard truth: Anthropic may be the tip of the spear in this moment, but AI capabilities in general and models from multiple companies and open-weight developers will almost certainly have similar capabilities to Mythos 5 in the near future -- if they don't already. "It's myopic in the extreme to think that no other competitors to Anthropic will develop similar capabilities to Mythos or even that they have not already done so," says Tarah Wheeler, chief security officer of the specialized cybersecurity consulting firm TPO Group. "There are other companies hot on Anthropic's heels who probably have the capabilities, too, and are holding them in reserve as they see how Anthropic is being treated in the current regulatory environment." Anthropic itself has emphasized this point since the launch of Mythos Preview. "The real message is that this is not about the model or Anthropic," Logan Graham, the company's frontier red team lead, told WIRED when Mythos Preview launched in April. "We need to prepare now for a world where these capabilities are broadly available in 6, 12, 24 months." OpenAI, for example, also did a private release of a cybersecurity-focused model in mid-April and announced an expanded cybersecurity strategy. Researchers note that even before this next generation of models, existing AI offerings could be used for advanced vulnerability-hunting and exploit development with a refined harness. A large group of cybersecurity leaders emphasized this to the administration in an open letter on Sunday, arguing that the White House's export-control directive was misguided. "It's not one model; it's the general trend of technology," says Bruce Schneier, a researcher at Harvard University and the University of Toronto who has been analyzing the situation. "Smaller, cheaper, open-source models, sometimes by themselves and sometimes in concert with each other, can match Mythos/Fable's performance with more sophisticated prompting. And we should expect other models to match Mythos/Fable's creativity and tenaciousness within months -- slightly longer for open-source models." What the White House and governments around the world need to focus on, experts say, is democratically developing much broader and more transparent plans for how they will contend with advances in AI capabilities on cybersecurity and in other sensitive areas as they inevitably occur. "The policy question is not whether a technology has risk," says Chris Wysopal, cofounder of the cloud security firm Veracode. "The question is whether a specific restriction meaningfully reduces that risk or whether it mainly slows down the people trying to make systems safer." This story originally appeared at wired.com.

Encryption, spyware, and now Mythos: History shows why cyber export control doesn't work

Last Friday, citing unspecified national security concerns, the White House ordered Anthropic to restrict the export of its powerful AI models Fable and Mythos to anyone outside of the United States, as well as foreign nationals inside the country. Shortly after, the AI giant hastily pulled the plug on both models, which have now been unavailable to anyone for a week. The episode is the first real test of whether the U.S. government can use export controls to contain frontier AI the way it has tried, with very uneven results, to contain encryption and spyware before it. And dramatic as it may sound, how this standoff gets resolved could shape not just Anthropic's access to foreign markets but the rulebook that other AI labs will have to build around. Some context first. Ever since Anthropic launched Mythos in April, the company has marketed it as some kind of Doomsday cyber machine that could wreak havoc on the internet if released too widely -- which is why, before the ban, only around 150 vetted companies and government organizations had access to it at all. The goal was helping defenders secure their software and services before the bad guys could reach Mythos-like capabilities. So what triggered the ban? Two subsequent events, reportedly. The first: Anthropic gave a South Korean telecom access to Mythos through its limited partner program, and U.S. officials grew alarmed after identifying the company as one they suspected had ties to China. (The company, widely reported to be SK Telecom, has denied any China connection.) Amazon CEO Andy Jassy also reportedly alerted the administration after Amazon's own researchers, he said, found a way around Fable 5's safeguards. Anthropic disputes the "jailbreak" label, calling it a narrow, already-patched issue rather than a wholesale defeat of the model's safety measures. The result was the same: the Commerce Department issued an export control directive, and Anthropic had to scramble to immediately limit access to its products within roughly 90 minutes of being notified, by some accounts. None of this is new, though. Governments have tried to use export controls to limit the proliferation of what they see as dangerous cyber technology for decades, but their track record has been middling at best. The U.S. government was behind what is perhaps history's most spectacular failure of this approach in the early to mid-1990s. At the time, computer scientists were developing encryption technologies to secure data as it traveled over the internet. One of those encryption products was called Pretty Good Privacy, or PGP, a popular software that could encrypt data and make it virtually impossible to unscramble even if intercepted as it traveled to its intended recipient over the internet. The U.S. government initially saw PGP as a dangerous weapon, fearing it would prevent its intelligence agencies from snooping on emails as they crossed their wires. To stop the distribution of PGP, the U.S. Customs Service opened a criminal investigation against PGP's creator Phil Zimmermann for allegedly violating arms export controls. He fought back by publishing PGP's source code as a printed book, igniting what is known today as the "Crypto Wars." Zimmermann later won a key battle when the investigation was closed, paving the way for crucial end-to-end encryption algorithms such as the one used by billions of Signal and WhatsApp users. Later during the early 2010s, researchers began discovering Western-made spyware used against dissidents in the Middle East. In response, several governments agreed to expand the Wassenaar Arrangement, an international treaty that limits the export of dual-use software and technologies that are used in both civilian and military applications. The idea was to classify surveillance and hacking software as dual-use, thus forcing spyware makers to get export licenses to sell their products abroad. But Wassenaar has always had two inherent weaknesses. There are several countries that don't adhere to the agreement, including Israel, which houses some of the world's most active spyware makers. The agreement also depends on countries applying it to companies within their borders at their own discretion. For a time, the Italian government allowed one of the country's then-top spyware makers, Hacking Team, a license to export its tools around the world, despite the company's track record of selling spyware to oppressive governments that used it to hack journalists and human rights activists. Since then, other countries in Europe have been lax with spyware makers like Italy. Despite numerous scandals, Europe, home to many spyware and hacking tools makers, has continually failed to curb the export of spyware to authoritarian regimes. Critics say that a recently renewed effort across the bloc of 27 member states to tackle its growing problem of spyware exports to authoritarian states "does not go far enough." Several spyware makers, such as Intellexa, a sanctioned consortium of spyware companies, have simply moved their operations to countries with lax export controls. Other spyware makers sought to move their operations to Saudi Arabia for similar reasons. There have been some wins. Germany-based spyware maker FinFisher shut down in 2022 after a multi-year investigation by German prosecutors into the company for allegedly selling spyware to Turkey without an export license. Investigators previously found the FinFisher spyware had been deployed on the phones of critics of Turkey's government. As of the time of writing, the impasse between Anthropic and the Trump administration remains. There is a reasonable chance the administration will buckle and lift the restriction in the interest of keeping American AI companies competitive worldwide -- a move that would amount to tacit acknowledgment that AI labs elsewhere, including in China, will likely reach similar capabilities regardless of what the U.S. restricts. Or, American AI companies could end up needing government approval before serving foreign customers at all, a compliance burden that would invariably dent their bottom line. Given the past experiences that world governments have had with trying to control the reach of software, government-mandated export controls are unlikely to be the right approach to stop malicious actors from abusing powerful dual-use cyber technologies.

US Limits on Anthropic Fable AI Could Hurt Cybersecurity

Fable 5 was built to help with advanced cybersecurity work. Its sudden shutdown highlights a dilemma at the heart of AI security: the same tools can aid both defenders and attackers When Anthropic released Claude Fable 5 on June 9, the company presented it as a new kind of high-powered artificial intelligence system: capable enough to help with advanced cybersecurity tasks but wrapped in safeguards meant to keep users from turning those same abilities toward attacks. Just days later, Fable 5 was offline. Citing national security concerns, the Trump administration ordered Anthropic to suspend foreign nationals' access to Fable 5 and Mythos 5, a closely related version with some safeguards lifted. Anthropic responded by disabling the models for all customers; the company said the order applied to any foreign national, whether inside the country or not, including those who are Anthropic employees. On supporting science journalism If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today. The move exposes a difficult trade-off at the center of AI security. The administration is trying to stop powerful models from being used by hostile actors, but the same systems are also becoming tools for defenders trying to find and fix weaknesses before attackers exploit them. If the most capable AI systems can be withdrawn overnight by political order, the result may not be a safer Internet but a more fragmented one, where companies and critical infrastructure providers, especially outside the U.S., are pushed toward weaker tools, rival Chinese systems or workarounds that are harder to monitor. Two recent cases show what defenders stand to lose. Mozilla has said its Firefox team used Claude Mythos Preview to identify 271 vulnerabilities that were later fixed in Firefox 150, and that earlier work with Anthropic's Opus 4.6 helped uncover 22 security-sensitive bugs in Firefox 148. Cloudflare has also said it tested Mythos against live code in critical parts of its infrastructure, where the model could chain lower-severity bugs and generate proof-of-concept code to test whether those vulnerabilities were exploitable. Peter Swire, a professor at the School of Cybersecurity and Privacy at the Georgia Institute of Technology and a former adviser to the Clinton and Obama administrations, says Anthropic's broader shutdown may have been precisely the outcome the Trump administration wanted. Although the order was framed as an export-control action aimed at foreign nationals, the company responded by pulling the models for everyone. "The administration used the legal tool of export controls, but their real goal was to block use by everyone, Americans included," Swire says. Swire sees an analogue in the 1990s fight over encryption. In the early part of that decade, U.S. rules treated encryption products as munitions and tightly restricted their export. The result was an awkward split between what could be used domestically and what could be shared abroad, before the Clinton administration loosened many of those controls later in the decade. Cybersecurity leaders are now making a related argument about AI. In an open letter, dozens of cybersecurity experts and executives urged the government to lift the restrictions. Cutting off access, they argued, could slow the people trying to find and fix software flaws before hackers exploit them. They have a point, Swire argues. "Going forward, banks and other critical infrastructure should be using the best AI to scan their own systems for defensive purposes," he says. And despite the Trump administration's "America First" policy, blocking access outside the U.S. could also harm the nation's security. "The U.S. hurts its own national security if the critical infrastructure of U.S. allies is undermined due to blockage of the best tools for defenders, so we end up in a less safe place," Swire says. In interconnected financial and software systems, a vulnerability in an overseas partner does not remain a foreign problem for long. Modern corporate networks, financial systems and software supply chains are interdependent, and attackers can enter U.S. systems through a weaker partner abroad and then move laterally. "If major European banks go down, that hurts the United States," Swire says. "U.S. cybersecurity defense depends on effective protections for all of their counterparties, many of whom are outside of the United States." Alan Woodward, a professor of cybersecurity at the University of Surrey in England, says the restriction is a "very blunt instrument." Anthropic has emphasized the power of its "Mythos-class" models, but Woodward says that message may now be working against the company. Woodward worries about the signal the restriction sends. "What they are going to do, of course, is put people off relying on U.S. companies," he says. "My fear, then, is that the Chinese will storm in -- and they already are, which we saw happen with DeepSeek -- and they'll capture the market, and they'll be controlling our use of it in a very different way." The question for Washington is whether it has made the dangerous use harder -- or simply made legitimate defense slower. As Woodward says, if you have an outright ban on something, "you effectively do lose control of it, because people will find other ways around it."

'Dangerous' AI Models Are Coming No Matter What

Late last week, Anthropic took its new Claude Fable 5 and Mythos 5 AI models offline following a United States government export-control directive barring "any foreign national" from using the services. The company has been in talks with the White House since Friday but has yet to secure an agreement that would allow it to reinstate the offerings. Since Mythos debuted in April, Anthropic has claimed -- and warned -- that the model has advanced capabilities for not only finding software vulnerabilities to help defenders patch them, but also figuring out ways to exploit them that could be used by bad actors. Anthropic itself noted this double edged sword in its launch of Mythos 5 and Claude Fable 5. "A great deal of advanced usage of AI models is dual use: the same queries that are beneficial in the hands of cybersecurity professionals and biology researchers could be dangerous if available to malicious actors," the company wrote in a blog post last week. With this in mind, the company initially released a version called Mythos Preview to a select consortium as part of a working group known as Project Glasswing. Mythos 5 was also privately released to this group last week, while Claude Fable 5, which is a Mythos-grade model, was released to the general public with specific blocks on its ability to give responses to questions about biology and cybersecurity. Then, at the end of last week, the Trump administration moved to restrict both models because it believes that Fable 5's guardrails can be disabled to allow full access to the Mythos 5 capabilities, allegedly making it a national security risk. Experts say, though, that this institutional clash is simply delaying or masking a hard truth: Anthropic may be the tip of the spear in this moment, but AI capabilities in general and models from multiple companies and open-weight developers will almost certainly have similar capabilities to Mythos 5 in the near future -- if they don't already. "It's myopic in the extreme to think that no other competitors to Anthropic will develop similar capabilities to Mythos or even that they have not already done so," says Tarah Wheeler, chief security officer of the specialized cybersecurity consulting firm TPO Group. "There are other companies hot on Anthropic's heels who probably have the capabilities, too, and are holding them in reserve as they see how Anthropic is being treated in the current regulatory environment." Anthropic itself has emphasized this point since the launch of Mythos Preview. "The real message is that this is not about the model or Anthropic," Logan Graham, the company's frontier red team lead, told WIRED when Mythos Preview launched in April. "We need to prepare now for a world where these capabilities are broadly available in 6, 12, 24 months." OpenAI, for example, also did a private release of a cybersecurity-focused model in mid-April and announced an expanded cybersecurity strategy. Researchers note that even before this next generation of models, existing AI offerings could be used for advanced vulnerability-hunting and exploit development with a refined harness. A large group of cybersecurity leaders emphasized this to the administration in an open letter on Sunday, arguing that the White House's export-control directive was misguided. "It's not one model; it's the general trend of technology," says Bruce Schneier, a researcher at Harvard University and the University of Toronto who has been analyzing the situation. "Smaller, cheaper, open-source models, sometimes by themselves and sometimes in concert with each other, can match Mythos/Fable's performance with more sophisticated prompting. And we should expect other models to match Mythos/Fable's creativity and tenaciousness within months -- slightly longer for open-source models." What the White House and governments around the world need to focus on, experts say, is democratically developing much broader and more transparent plans for how they will contend with advances in AI capabilities on cybersecurity and in other sensitive areas as they inevitably occur. "The policy question is not whether a technology has risk," says Chris Wysopal, cofounder of the cloud security firm Veracode. "The question is whether a specific restriction meaningfully reduces that risk or whether it mainly slows down the people trying to make systems safer."

Anthropic's latest feud with the Trump admin may actually help it, sales data suggests

Anthropic is having a month. The AI lab finished May by surpassing OpenAI in market share of business spending for the first time, Ramp just revealed. It raised $65 billion at a $965 billion valuation (also besting OpenAI) at the end of May, then waltzed into June by filing confidential paperwork for an IPO, reportedly on the strength of its first-ever profitable quarter. Then on Friday, the Trump administration renewed its war on the model maker by sending a letter demanding it ban non-Americans, including Anthropic's employees, from accessing its state-of-the-art models: the limited-release Mythos 5 and the more guarded version of Mythos released to the public three days earlier, called Fable 5. This essentially forced Anthropic to pull its latest all-powerful model from the market altogether. Although the White House invoked an obscure export control directive when ordering the ban, the exact cause remains unclear. The chatter was that hackers easily bypassed Fable 5's guardrails, which were intended to prevent access to Mythos' capabilities. That model is so good at finding security flaws in software code that Anthropic itself marketed it as dangerous and restricted its public release. This new drama comes after Anthropic famously refused to allow the government to use its models for mass surveillance of Americans and fully autonomous weapons. As a result, in March, the Trump administration declared the company a supply chain risk. That didn't deter Anthropic's sales to businesses. Quite the opposite, Ramp's data shows. Ironically, this latest feud with Trump administration, which also appears to validate the hubbub over Mythos' mythological power, may help rather than hurt Anthropic, according to Ramp's lead economist Ara Kharazian. Kharazian is the person who compiled the business spending AI data. "If anything, it'll probably boost them," Kharazian told TechCrunch. "Anthropic's best month on record, as far as business adoption, was the month that the Department of Defense labeled them a supply chain risk. There's a lot of aura that comes with your model specifically being named too dangerous to use." Ramp's data isn't granular enough for us to see how much of a financial hit the company will take by pulling Mythos and Fable 5 off the market. Still the data, from more than 70,000 businesses that use its platform, shows that customers heavily use Anthropic's Opus models and that business use has been growing. For instance, Ramp reported that Anthropic's share of AI subscriptions paid for by businesses rose 2.5 percentage points in May to 41%. This compares to OpenAI, which commanded 39.5% of AI subscriptions by its customers, essentially flat from the prior month. (OpenAI still greatly leads Anthropic in overall consumer usage, according to new data from Sensor Tower). Beyond subscriptions, the vast majority of what companies spend money on is API calls to the model, which cover token use for activities like coding. Anthropic's Claude Code has a strong reputation as a powerful AI coding tool. Ramp can't always see from the spending data which models most businesses are using. When it can see the model details -- in about one-third of transactions -- businesses are mostly spending on various flavors of Claude Opus, particularly the later versions. Opus is the model that preceded Mythos and is still openly available. In fact, in late May, Anthropic released a new version, Opus 4.8. Mythos had not been on the market for that long, having been released to limited users as of April. And Fable 5 was shut down after a few days. While we can't predict how this latest drama with the White House will impact Anthropic's ability to go public as it hoped to (public-market investors tend to be wary of companies embroiled in controversies with the government), the numbers indicate that Anthropic's available models are more popular with businesses than ever before.

Anthropic Is Still at Odds With the White House Over Claude Fable 5

Trump administration officials concluded talks with Anthropic on Monday without lifting export controls that were imposed last week on the company's most advanced AI models in response to jailbreaking concerns, according to three people briefed on the matter. The administration continues to believe that there are ways to disable some of the guardrails on Anthropic's Claude Fable 5, effectively allowing users to access the more powerful cybersecurity capabilities of the company's Mythos model, the people said. Anthropic has said for days that the administration's concerns are overblown, a position it reiterated in working group meetings held at the Commerce Department with government researchers from Center for AI Standards and Innovation (CAISI) and the Office of the National Cyber Director Sean Cairncross, one of the people said. The meetings were also attended by Commerce secretary Howard Lutnick, who dialed in by conference call from the G7 summit in Evian, France. Cairncross himself did not participate, the person said. On Anthropic's side, cofounder and chief compute officer Tom Brown and head of external affairs Sarah Heck have been leading the discussions. Anthropic's head of frontier red teaming, Logan Graham, and senior security researcher Nicholas Carlini flew to Washington, DC for the talks. "Both parties are working quickly to get this resolved," an Anthropic spokesperson said in a statement to WIRED. A White House spokesperson declined to comment. It was not immediately clear how any next steps might play out. The Commerce Department expressed a willingness to find a way to bring Fable 5 back online for consumer use, but it would likely be contingent on Anthropic fully resolving the jailbreak concerns, the person said. Ringing the Alarm The emergency talks have come at a fraught political moment for Anthropic, which was already in a prolonged fight with the Pentagon over whether its AI models could be used for certain military applications. The Trump administration was first alerted to the jailbreak concerns last week. Amazon CEO Andy Jassy called Treasury secretary Scott Bessent directly about the alleged vulnerabilities, which played a role in spooking the administration, the people said. Jassy's conversation with the Trump administration was first reported by The Information. Alarmed White House officials tasked the NSA to help review the vulnerabilities. The NSA responded that it believed it was indeed possible to strip away Fable 5's guardrails, prompting the administration to impose restrictions on the model. Lutnick then spoke with Anthropic chief executive Dario Amodei on Friday, as the Commerce Department drew up its letter imposing export controls on Fable 5. Over the weekend, after Anthropic cut off access to the model for all users, Lutnick was on multiple calls with Brown and Heck, according to a person with knowledge of the events. It's unclear why Amazon, one of the largest investors in Anthropic, rang the alarm on Fable 5. "As a leading cloud provider that serves a large number of private and public sector customers, it's not uncommon for governments to seek our counsel on potential security risks," an Amazon spokesperson tells WIRED. "When they occur, we don't share the details of these discussions." Security Disconnect At the core of the conversations between Anthropic and the administration is a disagreement over the severity of the Claude Fable 5 jailbreaking concerns. In a blog post on Friday, Anthropic implied that the administration's characterizations of the potential risks are overblown. Some cybersecurity researchers reiterated this position to officials on Monday, sending an open letter arguing that the export control action taken against Anthropic was unjustified. "Anthropic's Mythos-class models are quite good at finding flaws and weaponizing exploits. However, they are not uniquely good at these tasks, and many of the undersigned individuals regularly use other foundation and open-source models for security audits and red-teaming every day," the open letter reads. "As a result, this action has taken the best models away from defenders, created market uncertainty, and risked America's AI leadership without any real risk to justify it." Jailbreaking works by prompting an AI model in specific ways to circumnavigate its safeguards. Because Fable 5 is a version of Mythos with certain cybersecurity, biology, and chemistry guardrails in place, getting around those protections would effectively give users a version of Mythos. Anthropic has itself raised significant concerns about allowing Mythos to be used by the general public; however, it said on Friday that Fable 5's safeguards were strong enough to allow for a public release. Researchers who evaluated Amazon's findings say that the issues identified did not fully nullify Fable 5's safeguards. "It wasn't a jailbreak per se," says Katie Moussouris, founder and CEO of Luta Security, who published an analysis after reading the Amazon paper. Moussouris emphasizes that regardless of whether the US government has proof of a full Fable 5 jailbreak, restricting the model's ability to access to certain topics is a stopgap at best. "Most of us [in security research] think guardrails are speed bumps and shouldn't be treated like security boundaries for skilled adversaries," Moussouris says. "They only serve to slow down the less skilled." Investors in Anthropic have also been working over the weekend, trying to assess how the company's latest spat with the White House affects its corporate future, says another person close to the company. Some investors believe the US government is singling out Anthropic, and a competitor may not have faced the same reaction if they released a model similar to Mythos, the person says. The White House's export control directive also raises broader questions for other AI labs aiming to release AI models with Mythos-level capabilities, and how they can do so in compliance with the US government. It's now expected that AI labs give the White House early access to advanced AI models, and that they be extremely proactive about keeping the US government informed on model launches, according to AI lab leaders who spoke with WIRED. "The events over the weekend ... are informative for everyone that the [US] government would be willing to take these steps," says Aidan Gomez, CEO of Cohere, a smaller AI lab based in Canada that offers enterprise tools. "No one can be naive to that reality."

Anthropic got hit by export rules nobody understands

Anthropic has spent much of this week fighting to get its newest AI models back online after the Trump administration abruptly ordered the company to cut access for all foreign nationals, including users inside the US and its own employees, forcing Anthropic to block access to Fable 5 and Mythos 5 for everyone. The Trump administration has not publicly explained the legal basis for the order, but in a statement on its website, Anthropic said the government cited "national security authorities" to justify "an export control directive" on the models. (Anthropic also claimed that the government's concerns about a "jailbreak" potentially used by groups linked to China to access its models did not allow users to circumvent all of the company's safeguards.) But why did the administration use export control rules to address this? Experts say the episode appears to be unprecedented, exposing an uncertain and unstable stage in AI governance. And what, exactly, is Anthropic supposed to be exporting? (The company did not respond to The Verge's request for comment.) Export controls have traditionally applied to things that can be shipped across borders: weapons, hardware, tools, that kind of thing. Over time, the framework has expanded to cover less tangible goods, such as software, source code, technical data, and even 3D-printed gun files. These are still discrete things that can be copied, downloaded, published, or otherwise handed over and taken, not simply used through a remote service like a chatbot. In the context of AI, President Joe Biden moved to control AI model weights -- the core data that makes a model work that can be copied and run elsewhere -- in this manner; this idea was swiftly abandoned by the Trump administration in the second term. The Anthropic order does not fit neatly into this framework. There is no obvious transfer taking place: Mythos and Fable remain hosted on Anthropic's servers, and users do not receive source code, model weights, or a copy of the model themselves, instead getting the chatbot's responses to their queries. The export could be some specific information produced by the models, but it's not clear why that would require disabling access to the entire system rather than just restricting part of it. It could also be access itself -- though remote access to cloud services is a known gap in current export control regimes, one that Congress is already trying to close through legislation now moving through the Senate. Hanna Dohmen, a senior research analyst at Georgetown University's Center for Security and Emerging Technology, told The Verge it is "an open question" as to whether the order strains existing rules without seeing the precise language behind it. "In any case, this regulation is quite notable because, to my knowledge, this is the first time US export controls have been used to control access to an AI model in this way." "To say that this is an unsettled area of export control rule-making would be an understatement," said Andrew Reddie, a professor at UC Berkeley's Goldman School of Public Policy. He said that export control rules and other regimes like arms regulations give the government "wide latitude" to restrict access to certain goods. But "the equivocation by successive administrations regarding what the responsibilities of model developers are" has made it hard for firms to understand what is expected of them, he said. That leaves the industry in a bind. If Anthropic was targeted because Mythos and Fable are uniquely capable, the order raises obvious questions for the next generation of models from OpenAI, Google, Meta, xAI, and any other frontier lab. If they were targeted because of specific safeguard issues, the government needs to outline what protection it considers sufficient. And if Anthropic was singled out because of its testy relationship with the Trump administration, the order becomes even harder to make sense of. Either way, experts say this is not a sustainable way to manage frontier AI, especially if the US wants to maintain its lead globally. The incident has already added fuel to arguments that governments and companies outside the US should be wary about relying on American firms for access to strategically important systems. Reddie had similar concerns. "In some ways, I think this episode makes clear the unsustainability of the existing governance regime," he said. That is especially true if the government was more concerned about whether users could jailbreak models and bypass their safeguards. "If creating models that are impossible to jailbreak becomes the de facto standard for the United States, then it will have no AI models." All of this points to the same problem: The Trump administration wants it both ways on AI. It has repeatedly said it wants to take a hands-off approach and champion American technology, yet forced a domestic champion to unceremoniously yank its frontier models through an order it has still not publicly explained. If Washington wants to control who can access powerful AI systems, it needs to say how, and give companies an actual chance of complying before launch. Ad hoc interventions seemingly delivered on a whim are not sustainable in the long run -- and they are a good way to make sure the US falls behind in the AI race.

Total Recall: A Cautionary Fable Of Anthropic And The U.S Government

On Friday, June 12, the same model class covered by our previous blog post went dark. Anthropic suspended Fable 5 and Mythos 5 worldwide after a US Commerce Department issued an export-control directive which led to requests from prominent cybersecurity pros requests to undo the action. The bypass that triggered the export controls, per Katie Moussouris, was a request to "fix this code" that required a multi-step and manual process. This occurred after a request to "review this code for security issues" failed and triggered the guardrails. Our original launch analysis still holds, but the revocation adds a risk vector that isn't about the model's capabilities at all. It's about who can switch them off, how little warning you'll get, and the business risk implications for organizations. Once model access is restored, this post will remain relevant in case it happens again. Here's the "good news" (if there is any about this episode). Fable 5 (and its brother model Mythos 5) just released. While plenty of organizations had internal access for testing purposes, most enterprises had yet to build any meaningful integrations into existing products and services. Consumer subscribers had access to Fable 5 as well, but only with significant utilization premiums attached to it. The reality is that the revocation of access to these models came so quickly that it's unlikely it resulted in substantial impact to enterprise teams. That's the bright side. The downside, however, is that the next suspension may not come as quickly as this one. We dive into the details of what this means below. What The Revocation Means 1. It does not mean "go buy a non-US model", but it does mean single-sourcing any frontier model is now a concentration and continuity risk. The instinct to flee to a non-US provider is understandable and wrong. A European or Chinese model carries its own export-control exposure, its own government-control story, and in the China case a far more direct state-access concern for US enterprises. Swapping one sovereign dependency for another simply relocates risk under the guise of taking action. AI sovereignty is also not a "plug and play" choice of AI model alone. It also depends on factors like compute infrastructure, data sovereignty, and energy sources and non-technical inputs like operational, legal, and compliance risk when working in other jurisdictions. What the revocation actually argues for is model portability as a control objective: abstraction layers that let you reroute workloads across providers, fallback models tested before you need them, and an inventory of every workflow that hardcodes a single model string. Single sourcing is the issue, not US government oversight. However, the action itself should spur two discrete actions from organizations: 2. It does signal an era of direct government involvement in model releases. The June 2nd executive order established a voluntary framework for pre-release government access to frontier models. Ten days later, the US government pulled Fable 5 and Mythos 5 with an export control directive. That makes the "voluntary" part of the framework seem mandatory. Whatever the merits of this specific case, frontier models are being treated as objects of national-security policy, and "generally available" is now conditional. For US enterprises, this distinction can't be ignored: a line now exists between your AI vendor's product roadmap and US national security policy. Model availability is no longer purely a commercial SLA question. Since the US cited an available jailbreak for Fable 5 as the reason for the export control directive (which Forrester predicted would be developed in our first blog), this also means that model availability is now at the mercy of AI safety protocols and knowledge from users and security practitioners of available ways to bypass controls. Since AI models are probabilistic, not logical, jailbreaks are likely always possible meaning sliding them constantly in the crosshairs of political economic policies. 3. "Available in the cloud" is no longer the same as "available to you." Every assumption your teams hold about SaaS and API availability has changed. It's now true that a frontier model can disappear from your stack for reasons that have nothing to do with the vendor's competence or your contract, and the vendor may be legally barred from giving you notice. Add that to your risk register. The revocation also adds the inverse exposure: any vendor that already built on a model can have it pulled out from under them. Your third-party risk assessments need to capture not just which vendors use frontier models, but which models are embedded in each tool, automation, chatbot, or development workflow and whether those vendors have tested fallbacks of their own. A vendor with no Plan B for a model suspension now carries a continuity risk. Multi-model routing plans require engineering, architectural, and design revisions as model differences result in behavioral differences. Having a software inventory is the only way to take control of this. You need to know which third-party software incorporates AI models from AI coding agents and open-source build tools as well as what models AI-powered business applications and SaaS platforms use. This requires reviewing your vendor contract language today. Ask suppliers for a Software Bill of Materials (SBOM) that includes detailed information about the AI models they use. They may also include an AI-BOM. Talk to your suppliers about how they choose their models, when changes occur, and how you'll be notified. If you don't know what AI is embedded in the tools you buy, download, or use, you're exposing your organization to governance challenges that span security, IT, and the business. 4. Export-control now reaches inside your customer base. The directive reportedly restricts access for foreign nationals, including a vendor's own foreign-national employees, under "deemed export" logic. This could usher in the era of "know your customer" for AI access across enterprise and commercial subscriptions. This would mirror the KYC requirements originally issued in Executive Order 13984 that forced IaaS providers to verify the identities of foreign customers. Frontier model providers would need to collect citizenship details of employees and subscribers across their customer base to meet the export control requirements mandated by governments. For example, if Anthropic had this capability already, it would not have needed to suspend all access, merely prove that it could bar access to non-US citizens only and that its customers could do the same. Ensuring that Anthropic partners that host or resell the model along with their customers abided by the restriction is a bigger challenge. That has enormous impact on customer identity and access management (CIAM) and workforce identity. While controlling this at the workforce level is far easier and well understood, if frontier model access for customers starts depending on nationality, legal, privacy, HR, security, and IAM teams will have to agree on what data can be collected, who can see it, and how it's used. Connect With Us Forrester clients with questions related to this can connect with us through an inquiry or guidance session.

The US government's Anthropic models ban was never about an AI jailbreak

The U.S. government's enforcement letter to Anthropic, which effectively forced the company to pull its latest AI models offline just before the weekend, should be a wake-up call for any U.S. tech company -- AI lab or otherwise. To catch you up on the news blitz: On Friday afternoon, the U.S. Commerce Department sent Anthropic a letter invoking an obscure export control directive that banned non-Americans, including Anthropic's employees, from accessing Fable 5 and Mythos 5, citing an unspecified national security concern. Anthropic said it believes the letter is related to a bypass of the model's guardrails, but isn't sure because the letter doesn't provide specific details. The letter has not been made public. In response, Anthropic shut down both of its top models to all customers to ensure that it complied with the directive. The result was that the U.S. government successfully forced a tech company to pull its models offline with a swift and unilateral action that didn't appear to require court approval. Friday's intervention by the Trump administration shows that the AI industry is not immune to government interference. It's also a warning to the wider tech industry: comply, or we can shut you and your products down. Citing sources, Axios described a tense situation over the weekend between the two major players, saying that the "personality differences" between Anthropic and the Trump administration led to the export directive, rather than a technical issue with the AI products. New details about the issue that emerged over the weekend now cast further doubt on the government's already shaky reasoning. Katie Moussouris, a cybersecurity veteran and researcher who founded Luta Security, said in a blog post that Anthropic recently shared with her a private copy of a paper written by security researchers describing an alleged guardrail bypass in Fable 5. (The Wall Street Journal reports that the paper's authors are security researchers at Amazon.) Moussouris said that Anthropic reached out to ask for her take on the paper. Moussouris' blog post described how the researchers triggered the guardrail bypass, but said that the bypass itself "should never have triggered an export control." The difference is largely between asking an AI model to "review code for security issues" versus asking it to "fix this code." The end result is largely the same, even if the questions are posed slightly differently. "The behavior described in the paper cannot meaningfully be fixed, and any attempt would only weaken the model for defense," said Moussouris, who criticized the export control directive as hasty, heavy-handed, and misguided. Moussouris and dozens of other top security researchers and experts have since called on the Trump administration to revoke the export control order, calling the move to pull advanced cybersecurity capabilities from network defenders in the U.S. as "dangerous." Past administrations have made sweeping decisions on knowledge gaps. For instance, language used by the U.S. government during the 2010s to fix export law covering cybersecurity tools that could also be used for cyberattacks was so broad that it inadvertently near-outlawed legitimate security and vulnerability research. However, the Trump administration's directive appears retaliatory. Justin Hendrix, the editor of Tech Policy Press, said the Trump administration's move is "likely to raise alarms in foreign capitals about the reliability of American AI for critical applications." The message is that AI companies in the United States can't be trusted to operate without interference from the U.S. government. The Trump administration hasn't confirmed why it invoked its export control directive. Did the officials misread the report and freak out? Did Amazon CEO Andy Jassy say something to senior government officials that prompted the reaction, out of caution or spite? Was something lost in translation, or was this a way to pressure Anthropic, with whom the administration already has a fractious relationship? It's possible that the White House was unaware of the far-reaching consequences of the letter's demand and officials are scrambling to undo the damage of their own making. To quote Hendrix, "the climate is one of a cloud of suspicion that senior officials are picking favorites based on personal and political factors." The aftermath is that the government has set a dangerous precedent about how much control it intends to wield over the release of American-made software. This time the government took issue with Anthropic; tomorrow it could be with anyone else.

US pulls the 'kill-switch' on Anthropic's Fable 5 AI models, sending global allies scrambling -- European and Canadian leaders alarm allies over sudden export bans

World leaders are concerned that projects can be cut off by arbitrary U.S. regulations. World leaders have raised concerns over the U.S. administration's recent placement of export controls on Anthropic's frontier AI models, Mythos 5 and Fable 5, over national security concerns, Euronews reports. Suggesting this was a "wake-up call," moment, politicians and prominent figures in the UK, Canada, France, and the Netherlands, among others, said that frontier AI model access was now "critical infrastructure," and something that they desperately needed better control over. Many of them didn't even point at America directly, merely saying that if governments around the world can block access to the latest AI technologies arbitrarily, then it was within their national security interests to find alternative solutions. That said, that likely means building their own national AI efforts, fragmenting the industry, and reducing reliance and dependence on U.S.-based companies like OpenAI, Google, and Anthropic. Although Reuters reports the heads of U.S. technology firms like Nvidia and Adobe have been in talks with the Trump administration in the hopes that it will reinstate access to Fable 5 and Mythos 5, arguing that the bans hamper cybersecurity defensive efforts, the damage already appears done. The trust that some have had in access to U.S. frontier models is gone. The Myth, the Fable, the Cautionary Tale Anthropic debuted its 'game-changing' cybersecurity-focused AI model, Mythos, in April, claiming it was too dangerous to give the world wider access, but it brought in a few select companies and organizations under Project Glasswing to improve their code security. There was a lot of fearful marketing involved, but it was genuinely very good at finding flaws in old codebases. Anthropic suggested similarly capable models would be out in the wild within 18 months, so everyone needed to prepare. But in early June, it widened access to Mythos to 150 global organizations, and then a few days after that, dropped Fable 5, a Mythos-grade AI model, but with additional safeguards to protect against it being used for nefarious cybersecurity tasks. Despite those would-be protections, the U.S. government quickly swooped in and shut it down, claiming it had been jailbroken and was too dangerous to have in the wild. It placed export controls on the model, and by June 12, it was offline and inaccessible. On an individual level, the Claude subreddits have been filling up with programmers crossing their fingers that they'll be given access to Fable 5 again soon, but the more pronounced effect impacts global politics and national security. Wake Up Call Shutting down access to Fable and Mythos didn't just mess with programmer workflows. It shut down government and private projects all over the world, most of whom assumed that model access was all but guaranteed. Even if they didn't own the models, the free market would ensure they always had access to the best. But with the U.S. government's export block, that paradigm has shifted. "The United States is once again demonstrating what we Liberals and Democrats have warned about so many times since Trump entered into office; that the US holds a real 'kill-switch' over essential technologies and that they are more than willing to use it," said French Member of European Parliament, Christophe Grudler, in a statement. The concern over the U.S. government having too great a control of frontier AI model access is also leading to calls for Europe to develop its own alternatives, eschewing the need for American company technologies as much as possible. "These restrictions are a clear example of the current American 'nobody but us' mentality," said Dutch Renew Europe MEP Bart Groothuis. "Once again: this shows that Europe needs its own LLM's and open weight models or face digital colonization." Not every leader has been so pointed in their criticism of America. Canada's PM, Mark Carney, made it clear in his statement that "Nobody has done anything wrong in the situation," he said via APNews. However, he warned that "We will have done something wrong if we just accept this, don't take the lesson, don't build out and diversify." The UK's former minister for the Armed Forces and Labor MP, Al Carns, suggested this was just another example of why the UK needed to develop its own cutting-edge AI tools, leveraging its deep expertise in the field to ensure UK sovereign access to the most capable technologies. Some leaders aren't quite set on going it alone, though. France's President Emmanuel Macron championed a joint French and Indian AI effort. Speaking at an event in Nice on Sunday, Macon said: "Our two countries share the definition of a reliable, open and safe AI, that could be trusted, that could be responsible, that could be ethical," he said, via TribuneIndia. U.S. companies are scrambling for alternatives, too. Alex Stamos, CSO at Corridor, told The Verge that companies are rushing to sign backup contracts with non-US companies with open weight models so they can continue their projects undeterred, no matter what the Trump administration does next. In every instance, though, whether leaders pointed fingers or talked up their own efforts, wanted to go it alone or with new partners, the one clear dividing line is that not all of them are looking to move away from America. Alongside a number of other industries impacted by the Trump administration's tariffs and export controls, global partners that once saw the U.S. as the most reliable global partner are increasingly looking elsewhere as that evaporates. From AI, to Jets, to Search The U.S. government cutting off access to Anthropic's Frontier models happened quickly, and the consequences of the lost trust are likely to extend for years, or even decades, and affect far more than chips and models. Citing the recent case of Anthropic model access being pulled, France has announced it is switching from using a U.S. data and analytics firm, Palantir, for a domestic alternative, as Reuters reports. France is also transitioning government departments away from using U.S.-based messaging apps like WhatsApp, with a national alternative, according to Le Monde, Wired also highlights a number of instances of EU governments and organizations shifting away from U.S. tech firms, including changing default search engines from Google to Qwant, a move towards open-source office software developed in the EU over Microsoft and Google options, and many are ditching Amazon AWS and other U.S. cloud services. This recent Fable 5 shuttering is likely to only accelerate these efforts, as the reliability of access is called into question once again. But unraveling the EU and the rest of the world from America won't be easy, or even achievable, even in the long term. The global economy is still too integrated for that to be truly viable. But the desire and impetus are there. For key industries that impact national security - and AI alongside chip fabrication are becoming clear pillars in that space - national alternatives seem all-but-necessary for major militaries and economies. Whether that creates a multi-polar AI world, or just cements the clear headstart and advantage held by countries like the U.S. and China, remains to be seen.

Sovereign AI? Anthropic shutdown reveals Canada's weakness

The United States government recently ordered AI company Anthropic to suspend foreign nationals' access to Fable 5 and Mythos 5, two of its most advanced AI models, citing national security concerns. Anthropic responded by disabling the models for all customers. Organizations in Canada, Europe and around the world that had embedded those tools in their workflows found them simply gone. No appeal process. No migration window. No warning. No jurisdiction over this decision. As the G7 summit wraps up in Evian, France, the Anthropic shutdown has put AI sovereignty and concerns about U.S. dominance high on the agenda. For Canada, the Anthropic shutdown is not just a technology story. It is the bill arriving for past choices -- years spent celebrating AI research leadership while under-investing in the commercial, capital and governance conditions to turn leadership into lasting sovereignty. Canada's new AI for All strategy recognizes that dependence on foreign technology is a strategic vulnerability. It commits billions to compute infrastructure, skills development and domestic AI capacity. These investments are necessary and welcome. But you can build your own data centres and still depend on someone else's models. Canada needs to build the commercial conditions, long-term investment, procurement pathways and scale-up support that would allow AI companies to grow here rather than leave. Canada's AI research leadership Canada's contribution to modern AI is a matter of historical record. Yoshua Bengio, a computer science professor at the Université de Montréal, helped make Montréal one of the world's leading AI research hubs. Computer science professor Geoffrey Hinton's foundational deep learning work at the University of Toronto, which led to a Nobel Prize, was supported by public funding and helped lay the ground for today's AI systems. In 2013, Google acquired Hinton's startup DNNresearch, which he had incorporated with two graduate students, Alex Krizhevsky and Ilya Sutskever. Both moved to Google. Later, Sutskever co-founded OpenAI, a leading player in the current AI boom. Both Krizhevsky and Sutskever were trained in a Canadian research ecosystem supported by public investment, and both went on to do their most consequential commercial work elsewhere. Canadian researchers remain deeply woven into the systems shaping the global economy. But there is a real gap between scientific leadership and commercial control, and Canada has not closed it. Sovereignty means access The Pan-Canadian AI Strategy, launched in 2017 with $125 million, was celebrated internationally as visionary. But Canada has struggled to turn that research strength into domestic commercial control. Only seven per cent of the IP rights generated through the strategy are owned by Canadian private sector firms. In 2024, two-thirds of high-potential Canadian-led startups that raised more than $1 million were headquartered outside Canada. And Canada deployed less than two per cent of global AI venture capital despite hosting roughly 10 per cent of the world's top-tier AI researchers. Research leadership alone does not create sovereignty. Sovereignty depends on who owns the companies, controls the models and sets the terms of access. And Canada is still playing catch-up. Strategic dependency Most Canadian organizations using AI license access from a small number of American companies operating under U.S. law and subject to U.S. government decisions. That relationship looks commercial until a national security order turns it into a strategic dependency overnight. Think of the Hydro-Québec scenario in reverse. Imagine Ottawa ordering Hydro-Québec to stop exporting electricity to the northeastern United States overnight on national security grounds. American businesses and institutions would quickly discover that a routine commercial relationship was also a critical dependency. Canada is now on the receiving end of that lesson with AI. Who decides access? The most advanced AI systems are not nuclear weapons. But they are becoming consequential enough to raise a similar governance question: when a technology can affect economic competitiveness, research capacity, public services and national security across borders, should access be decided by one government and a handful of companies? For now, that is largely how the system works. The U.S. controls an estimated 74 per cent of global AI supercomputer capacity (China controls 14 per cent), while leading American firms build and operate most of the frontier models others rely on. Questions of access, ethics and safety have global consequences, yet they are currently addressed largely through national laws and corporate decisions, and with no genuine consultation or public participation. The Anthropic case makes these questions harder to ignore. A board-level risk National strategies take years. The Anthropic suspension took hours. That is the kind of disruption organizations should plan for. Guidance from the U.S. Department of Commerce already recommends planning for deactivation or disengagement, including workarounds, alternative processes and clear communication when access changes. Canadian organizations need to treat AI dependency as a significant risk, not an IT purchasing question. That means knowing where AI is embedded in critical workflows, whether those workflows depend on a single provider, what happens if access changes and whether the organization retains enough human expertise to operate without the tool. It means demanding vendor contracts that specify what happens when access is suspended, what data is retained and what notice obligations exist. Canada's future voice The lesson is not to stop using AI. Adoption will continue and the productivity gains are real. The lesson is to stop assuming access will always be there. Canada is now investing billions to strengthen domestic AI capacity. That matters. But sovereignty is not simply about where the servers sit. It is about who controls the capabilities running on them and what happens when that control is exercised by someone else. Canada helped build the present and the future of AI. The question now is whether it is willing to build enough commercial and governance capacity to have a meaningful voice in who controls it. Because whether we have a say or not, we will feel the consequences.

US clampdown on Anthropic models sends EU sovereignty surge into overdrive

As Anthropic execs prepare to visit the White House after effectively being ordered to cease offering the company's Mythos 5 and Fable 5 models, the European Commission says the incident is another example of why the EU must achieve technological autonomy. Anthropic announced on Friday that the US government issued an export control directive that required the AI upstart to prevent any non-US citizens from accessing its cybersecurity models Mythos 5 and Fable 5. The order meant even some Anthropic staff could not use its models. And as there's no way to tell if someone on the internet is a US citizen, the order effectively meant that the AI company had to stop making the models available to everyone to ensure compliance. Anthropic isn't sure why the White House issued the order. "Our understanding is that the government believes it has become aware of a method of bypassing, or 'jailbreaking,' Fable 5," the company said. "To date, the government has only given us verbal evidence of a potential narrow, non-universal jailbreak, which essentially consists of asking the model to read a specific codebase and fix any software flaws. "Our understanding is that one potential jailbreak was shared with the government." The Wall Street Journal reports that the directive was the result of conversations held between Amazon CEO Andy Jassy and US officials, including Treasury secretary Scott Bessent, and Jassy's report of a possible jailbreak. Anthropic executives are set to meet with US officials at the White House this week to gain a fuller understanding of the developments that informed the directive, according to Axios. Whatever the Trump administration's reason for the order, Mythos and Fable remain unavailable at the time of writing. A case study for sovereignty The incident has not gone unnoticed. Thomas Regnier, spokesperson for the European Commission, said the body is still examining the directive's implications for the EU amid concerns that the US can switch off access to technology that allied partners could soon come to rely on heavily. "The Commission has taken note of Anthropic's statement regarding the US export control directive on its most advanced models and is assessing its implications, including for users in the European Union," he said. "We are seeing a new generation of highly capable AI models reach the market. These models offer significant benefits, including for cyber-defence, but they also raise serious cybersecurity concerns that need to be addressed. "This is a shared challenge, not one confined to a single jurisdiction or company. We believe that contingency measures taken in this light should not be discriminatory against partners. "This development is a further illustration of why Europe needs to strengthen its technological sovereignty, and it underlines the relevance of the cybersecurity and AI legislation already in place at EU level, including the AI Act, the Cyber Resilience Act, and the NIS2 Directive - as tools to manage exactly this kind of risk on our own terms. "We are looking closely at the practical consequences of this for European users of these services." The comments come days after the EU launched its European Technological Sovereignty Package, a slew of measures aimed at sharply reducing its reliance on technology developed by the US and China. Cybersecurity-specific AI models such as Mythos 5, Fable 5, and OpenAI's GPT-5.5 are still very early in their development, and are not yet available to many organizations, let alone casual users. The cost of dependency stays invisible until it's too late The US directive to prevent foreign nationals from accessing Anthropic's models will nevertheless prompt concerns among global partners and organizations about how a foreign government can simply revoke access to technology on which they may become highly reliant in the future. For Aled Lloyd Owen, chief of staff at Responsible AI UK, the news of Anthropic restricting access to its models only strengthens the case for the EU's plans to loosen its ties to US tech. "This is another incident that just proves the rule and proves that [the EU] must move faster and deeper, and really establish that independence as soon as possible," he told The Register. As for alternatives, Mistral AI is one of the EU's flagship AI development projects. It is widely regarded as a fast, capable, open-source model, but one that lacks the performance of "frontier" models such as those made by Anthropic and OpenAI. Owen said there is a limit to how quickly the EU can achieve autonomy, but the latest Anthropic story is "quite helpful in a lot of ways." "It's saying: 'You can't, from a commercial point of view, trust these bodies,' so to some extent, are you willing to sacrifice performance, both perceived and real, for European homegrown models that are not quite there but are certainly driving in that direction, in order to have a more reliable sovereign service? "So, the ability to shift is both technological, in terms of building effective models and building effective infrastructure, but will also involve weaning European companies from the high-capability overseas models that they're already using." Kate Hanaghan, chief research officer at TechMarketView, said: "Last week, I was talking to a couple of European integrators about exactly this issue. One framed it as 'The cost of dependency stays invisible until it's too late.' "For UK enterprises, the risk is now very clear. Depending on a single US frontier provider leaves operations exposed if that access is withdrawn. And this weekend showed it can happen without warning. Ultimately, that leaves Europe to work out what it should, and realistically can, develop for itself." Voices in the UK echo those in the EU. Kanishka Narayan, minister for AI and online safety, posted on X: "The main lesson: as we debate the future of national security and technological sovereignty, access to AI capabilities is crucial." I care about sovereign AI because it now decides our security Separately, he said: "We treat every other threat to our sovereignty with deadly seriousness, but we haven't learned to treat this one in the same way." "I care about sovereign AI because it now decides our security... it will reshape our economy faster than anything else we've seen in our lifetimes," he added. The MP went on to say: "I'm not going to pretend there's a simple switch that we can pull. There isn't. Britain needs more AI capability. This is the central political question of our time, and our first duty is to see it clearly before someone else decides the answer for us." Policy on the run The order has also angered others, for different reasons. A group of 54 security and AI experts co-signed an open letter to the US government after the directive was issued, calling on the government to lift the restrictions. They also asked the government to commit to a more transparent approach to handling AI risk assessments in the future, saying that it should be a more democratic process. Not all the signatories believe the US should have regulatory control over AI models (Anthropic believes the US rightfully holds the authority to block releases), but they said that materially impactful decisions should be grounded in science and security teams should be given time to prepare. The letter pointed out that vulnerability researchers and red teams are already relying on these models every day, and decisions to revoke access to them should be made through a democratic process, and should restrict capabilities only to the minimal extent necessary. "As a result, this action has taken the best models away from defenders, created market uncertainty, and risked America's AI leadership without any real risk to justify it," the signatories wrote. Who's next? In its response to the White House order, Anthropic asserted the allegedly problematic features of Fable and Mythos are also present in other models, including GPT-5.5. Anthropic has stated from the launch of Fable 5 that it believes developing AI models with perfect jailbreak resistance "does not appear to be possible today," and that no one has developed a universal jailbreak for its models to the best of its knowledge. It has long advocated for and continues to stand by its defense-in-depth approach to managing risks. ®

Inside the fight over Claude Mythos 5

As the rest of the country celebrated the USA's first World Cup win and the New York Knicks championship, Anthropic spent its weekend fighting the Trump administration over its latest model release. At 5:21 PM on Friday, the company received a US export control directive to suspend access to its Mythos 5 and Fable 5 AI models by "any foreign national" inside or outside the US, "including foreign national Anthropic employees." The only way that was possible, Anthropic determined, was to completely disable products it spent the past week hyping -- and travel to Washington, DC in hopes of changing President Donald Trump's mind. Now, over the coming days, the US government could dramatically alter the trajectory of the entire industry, dealing a major blow to American AI companies. Claude Mythos 5 and Fable 5 are built on the same foundation as Anthropic's Mythos Preview, which Anthropic dubbed too dangerous to publicly release. (The company's warnings could be seen as genuine concern or more hype for their own model -- or both.) Mythos 5 was made available to a select group of government agencies and companies, while Fable 5, which featured additional safeguards, was deemed "safe for general use." But when a report indicated those guardrails may have failed, Anthropic's dire warnings about Mythos falling into the wrong hands came back to haunt it. A source familiar with the situation, who participated in the negotiations between Anthropic and the Trump administration, said the administration called the AI lab on Friday around 1pm ET and gave the company a 90-minute ultimatum to shut down access to Mythos 5 and Fable 5. If it didn't, then the government would impose export controls on Anthropic by authority of the US Commerce Department. The source said that Anthropic executives were talking to the White House within 15 minutes of that first call, confirming that CEO Dario Amodei joined the discussions about an hour and 15 minutes after that initial call. Amodei directly spoke with US Treasury Secretary Scott Bessent, Commerce Secretary Howard Lutnick, and National Cyber Director Sean Cairncross, in some cases more than once, the source confirmed. Anthropic wrote in a release on Friday that the company believed that the government "believes it has become aware of a method of bypassing, or 'jailbreaking' Fable 5." Rather than an existential threat, though, Anthropic said that the jailbreak in question was a "potential narrow, non-universal" one that was "shared with the government" by an entity the company declined to name. Moreover, Anthropic said the behavior wasn't unique to Fable 5. "We have reviewed a report that we believe is the basis of the government's directive and validated that the level of capability displayed there is widely available from other models (including OpenAI's GPT-5.5)," Anthropic wrote. Semafor reported, citing one source familiar, that the hubbub began because the US government was concerned that a China-linked group had accessed the technology. But the source said that the China rumors went back weeks, referring to a large global telecommunications company that was initially cleared to be included in access to Mythos Preview, and that when the US government shared its concerns, Anthropic immediately revoked access. An X post by David Sacks, the US government's former AI and crypto czar who stepped down in March, didn't mention China either. Sacks did, however, mention the unnamed entity that had exposed the issue to the government, calling it "a highly credible trusted partner of both Anthropic and the USG who was testing Fable [which] came forward with a jailbreak of those guardrails." Some reports point to Amazon CEO Andy Jassy as the person who flagged concerns to the US government after researchers at Amazon had red-teamed Fable 5. That conclusion stands at odds with some independent red-teamers, who have said they were impressed with the level of the protections. The source familiar with the negotiations said that the Amazon research was explicitly mentioned in conversations with the US government. The person added that Anthropic had had access to that paper within days of the Friday export control directive and had been going back-and-forth since then with Amazon researchers to discuss it. Everything in that paper, the source said, could be achieved by OpenAI's GPT-5.5. Anthropic spent the weekend scrambling to make nice with the Trump administration, beginning with virtual meetings and then flying employees to DC, including Dave Orr, Anthropic's head of safeguards; Logan Graham, who runs its frontier red team and has led work on Project Glasswing; and Nicholas Carlini, a leading frontier developer and cybersecurity researcher. Axios reported, citing a source familiar with the Trump administration's thinking, that the company simply has repeatedly made missteps in its communication with the administration and that it "has not done a great job at trying to speak to the administration and appreciate the ideological differences." For Anthropic, the timing couldn't be worse: the company had banked on Mythos to help it recover, in part, from months of high-profile clashes with the US Department of Defense. The source familiar with the negotiations said that Anthropic pre-briefed the administration on Fable 5, and that the US Department of Commerce conducted testing pre-deployment, with no concerns shared at the time. The source added that Anthropic had been working closely with government agencies since Mythos Preview's release. The Trump administration initially took a hands-off approach to AI safety -- but post-Mythos, it has become more ambivalent, even as it frets over the threat of losing the AI race to China. Now, prominent cybersecurity leaders have warned that sidelining Mythos 5 and Fable 5 could give China a significant AI advantage. Trump's move has galvanized international calls for alternatives to American AI systems, while effectively putting a major US AI company's new flagship model on ice. A public letter from tech and cybersecurity executives called for restrictions on Fable 5 to be repealed on Sunday. "Not all of us agree that AI regulation is the right way forward," the letter states, adding that if regulations are going to happen regardless, then they should be rooted in "scientific evaluations developed with input from industry and academia." Alex Stamos, chief product officer at Corridor, told The Verge he organized the public letter because the countless number of vulnerabilities in the past decade-plus, written in a variety of different coding languages, require AI to patch before bad actors find them. "We're in a race, and I think policymakers don't understand that," Stamos said. "There's this weird arrogance, this idea that American labs are hugely ahead of our adversaries that will always be true, that it's really important to restrict access because of that. I just think that's foolish. If the labs are ahead, it's only by a matter of months. And you can see that in the open evaluations. The cutting-edge models are only something like six months ahead of the Chinese models -- and those are the models we know about." The public letter goes on to state that though Anthropic's Mythos-class models are skilled at finding cybersecurity vulnerabilities and taking advantage of exploits, they aren't "uniquely good" at these tasks and that Fable 5's safeguards "were so aggressive as to be the source of humor in the cyber community on launch day." Stamos told The Verge that "there's a real overstatement of Mythos' capabilities. Anthropic is somewhat responsible for this themselves, clearly ... Mythos is great, but the real turning point was really last year." Stamos said the industry is awash with backup contracts being signed with non-US companies and open-weight models being deployed on alternative hardware arrangements because the past weekend made political risk part of companies' business plans more than ever before. "They are laughing at us in Beijing right now," Stamos said. "One of America's champions is being kneecapped by the US government while we're in a race with the Chinese. It's just incredibly stupid. That's why I wrote the letter, and I think that's why a lot of people signed onto it." Ben Van Roo, co-founder and CEO of Legion Intelligence, a system of agents for the national security community, told The Verge that "the directive of 'no foreign national should use this model' is the most impossible thing to enforce." He added, "When I first read that, my whole... [network of] AI community nerds was exploding." To make matters even more urgent, OpenAI, Google, and Microsoft have all come out with their own comparable products to Anthropic's Mythos, making many of the same claims about their effectiveness and risks. If the Trump administration bans Anthropic's advanced cybersecurity models, it can make a case for banning its competitors' models, too. That could spur AI industry leaders to unite and help out Anthropic or, as with its fight over autonomous weapons with the Pentagon, position themselves as a safer and more compliant alternative. Even as the Trump administration is trying to free tech companies of regulatory hassles, the Anthropic order could amount to a dramatic restriction on powerful AI models -- depending on how the next few days play out. Legion Intelligence's Van Roo called it "uncharted territory" in the regulatory setting, adding that he doesn't think this is the last time something like this will happen. We've also entered the era of AI populism, when a growing number of people are pushing back against the AI industry's outsized influence and the concentration of power at the top via data center protests, pledges to quit using AI chatbots, lawsuits over wrongful deaths, and even attempted attacks on AI company CEOs. Van Roo says the Trump administration's recent moves against Anthropic could stoke "greater fears and concerns, potentially for the wrong reasons." The source familiar with the negotiations described the weekend's conversations as constructive, with some members of the administration admitting that putting export controls on model providers isn't ideal, since competitors with similar products may find themselves under the same restrictions -- and since the US government is currently exploring a program that would encourage the export of American AI systems. As Anthropic continues to negotiate with the US government, there's little chance that the company's other myriad issues with the Pentagon won't come up -- namely, the ongoing battle between Anthropic and the Department of Defense over acceptable usage policies for Anthropic's tech by the US military. "This is new and we've never had anything potentially this drastic before, and it does have some real ramifications" in terms of how to enforce access to powerful models, Van Roo said. "Who gets to use this new technology that continues to outpace our own ability to regulate it?"

Anthropic asked for regulation. Washington went much further

Anthropic said on Friday it received an export control directive, ordering the company to suspend access to its latest Claude models, Fable 5 and Mythos 5. Anthropic's longstanding push for regulation is coming back to bite it. The artificial intelligence company, now valued at close to $1 trillion, has spent years touting its dedication to safety. But for the second time this year, Anthropic has found itself caught in the Trump administration's crosshairs, this time out of concern for the safety of its newest models. Late Friday, a couple hours after SpaceX wrapped up its first day of trading following a record IPO, Anthropic said it received an export control directive, ordering the company to suspend access to its latest Claude models, Fable 5 and Mythos 5, "by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees." The administration cited "national security authorities" but didn't specify its concern, Anthropic said. The directive landed just days after Anthropic CEO Dario Amodei published an essay advocating for more "serious and binding regulation of AI," including the ability to block models if they are deemed unsafe. "Frontier AI models, like airplanes, should be required to go through technical testing and auditing, and their release should be blocked or reversed as a threat to public safety if they do not meet high standards of safety," Amodei wrote. It was the latest public statement from Anthropic encouraging greater government oversight of the rapidly evolving AI industry. Since splitting from OpenAI to start Anthropic in 2021, Amodei and his top executives have been staunch advocates for AI regulation and have supported legislation at both state and federal levels. The company lauded an AI executive order that President Donald Trump signed earlier this month as an "important step." But Friday's action wasn't the kind of oversight Anthropic had in mind. The company said in response that it disagreed that the Trump administration's finding was cause for a recall, and it characterized the disruption as a "misunderstanding." "As we have stated publicly, we believe the government should have the ability to block unsafe deployments, as part of a statutory process that is transparent, fair, clear, and grounded in technical facts," Anthropic said in a blog post. "This action does not adhere to those principles." Senior Anthropic employees flew to Washington, D.C., to meet with members of the Trump administration on Monday. The company told CNBC that "both parties are working quickly to get this resolved." Anthropic hasn't said when it expects its models to come back online. Trump's executive order, 10 days ahead of the directive against Anthropic, was thin on specific details. It asked companies, on a voluntary basis, to provide models to the federal government to assess their capabilities ahead of a full release. The order also gave administration officials 60 days to develop and maintain the relevant review frameworks for AI companies to consider. There's no indication that the forced suspension of Anthropic's latest models was tied to the executive order. Rather, according to The Wall Street Journal, the move was prompted by conversations that Amazon CEO Andy Jassy had with U.S. officials, including Treasury Secretary Scott Bessent. Citing people familiar with the matter, the Journal said Jassy told the administration that Amazon researchers used a series of prompts to get Fable 5 to serve up information that could aid cyberattacks. While the president's executive order called for voluntary adherence, Friday's directive had a very different tone, said Daniel Remler, a senior fellow at the Center for a New American Security. "This sure looks mandatory if there are going to be consequences for not doing what the government says," Remler told CNBC in an interview. Fable 5 and Mythos 5 built on the April release of Claude Mythos Preview, a powerful offering that excels at identifying security vulnerabilities within software. Anthropic limited the rollout to a select group of companies as part of a cybersecurity initiative called Project Glasswing. Mythos 5 was still limited to a select group of users, but Anthropic made Fable 5 available to its enterprise customers and paid subscribers. The company said the broad release was possible because of new safeguards that block responses in specific high-risk areas, including cybersecurity and biology. Anthropic worked with government agencies to test the models ahead of the release and received approval to deploy them, according to a person familiar with the discussions who asked not to be named in order to discuss confidential matters. There was no communication from the government of a national security threat, the person said. "To go from an environment in which we had no controls on model access at all to basically using this tool to take down one company's model in a few short hours is pretty remarkable," Remler said. He added that, given the volatile policy environment, it's the kind of "reactive regulation" the company should probably have anticipated. "It's hard to manage risks with a technology that's changing so quickly all the time," he said. It's all happening at a pivotal moment for Anthropic. The company and its chief rival, OpenAI, both confidentially filed their IPO prospectuses recently, setting up potentially historic share sales for investors eager to jump into AI. SpaceX's three-day rally following its IPO is seen by many as evidence of Wall Street's enthusiasm for new and big opportunities. Amodei has some loud voices on his side. Alex Stamos, chief product officer at Corridor and the former security chief at Facebook, penned an open letter on "transparent AI cyber protections," addressed to Commerce Secretary Howard Lutnick and National Cyber Director Sean Cairncross. The letter, signed by more than 150 executives and technical leaders, asks for the export control directive to be lifted. Stamos told CNBC in an interview that rules need to be "based upon science." "Those rules need to be written down and transparent. That has not happened," Stamos said. "There's nothing that Anthropic or anybody else can look at to say what can and can't I do." He added that for every American company, "the fear is that at any moment you can now, if you run afoul of the administration, be shut down for a completely arbitrary decision capriciously." For Anthropic, clashes with the government have become a big part of the company's story. Earlier this year, it engaged in a high-profile dispute with the Department of Defense. The DOD declared Anthropic a supply chain risk in March, a designation that requires defense contractors to certify that they will not use Claude models in their work with the military. In a post on X on Saturday, Defense Secretary Pete Hegseth addressed the government's latest directive, writing that "every passing day" proves why blacklisting Anthropic was "the right move." Anthropic sued the Trump administration in an effort to reverse the supply chain risk designation. That litigation is ongoing. Stamos said the administration clearly has a problem with Anthropic, one of the few notable companies to push back on its policies. "If they were fairly enforcing these rules, they would have to enforce them against OpenAI and Google as well," he said. One vocal critic of Anthropic has been David Sacks, a venture capitalist who formerly served as Trump's AI and crypto czar. After a company executive published an essay in October on risks of AI, Sacks accused Anthropic in a post on X of "running a sophisticated regulatory capture strategy based on fear-mongering." Sacks weighed in after Anthropic received the export control directive, writing on X that the "minimizing language" in the company's blog post is not in line with its brand as "the AI safety company." "The Admin values Anthropic's technical capabilities and feels that this issue, while serious, should be easily resolved," Sacks wrote. "The ball is in Anthropic's court." -- CNBC's Kate Rooney contributed to this report Choose CNBC as your preferred source on Google and never miss a moment from the most trusted name in business news.

As Anthropic suspends access to new models, India debates its AI future

Anthropic's sudden move to suspend access to its newest AI models following a U.S. government directive has raised fresh questions across the global technology industry. In India, the decision has reignited a long-running debate over whether one of the world's largest AI markets can afford to rely on technologies built and controlled elsewhere. The announcement came late Friday, when Anthropic said it had received the U.S. government directive requiring it to suspend access to its recently launched Fable 5 and Mythos 5 models for all foreign nationals, including its own foreign national employees. The move came shortly after the company announced a partnership with Indian IT services giant Tata Consultancy Services to expand enterprise AI adoption in India, underlining how closely the country's AI ambitions have become tied to technologies developed and governed in the U.S. While the broader implications remain unclear, some reports said the initial security concerns were first reported to the government by Amazon CEO Andy Jassy. And The Information said the White House is unlikely to extend similar restrictions to other AI companies and is privately blaming Anthropic's handling of alleged jailbreak vulnerabilities. Anthropic has disputed the government's characterization and argued the action should not have been taken. Regardless, the development has triggered debate among Indian founders, investors, and policy experts over whether the country should accelerate efforts to build domestic AI capabilities, deepen investment in open-source alternatives, or continue relying on a handful of U.S. frontier model providers. For some, the episode is a wake-up call on technological dependence. For others, it is a reminder that access to increasingly critical AI systems can be shaped by geopolitical decisions beyond India's control. India has become one of the most important markets for frontier AI companies. Anthropic and OpenAI have both described the South Asian nation as their second-largest market after the U.S., reflecting its growing importance in the global AI race. The companies have already set up their offices in India, expanded local hiring, partnerships, and enterprise initiatives in recent months, betting on India's vast base of developers, startups, and businesses to accelerate adoption of their latest technologies. For many in India's technology sector, Anthropic's Friday announcement was about more than just one AI company. It reopened questions about the country's long-term AI strategy and whether India could afford to remain dependent on a small number of foreign frontier AI providers. "It completely changes things," said Aakrit Vaish, founder of Indian AI venture platform Activate, referring to Anthropic's decision. "I think this materially changes the way all of us should be thinking about sovereign AI in India." Vaish told TechCrunch that he woke up on Saturday morning "shocked and confused" by the announcement and said it strengthened the case for developing domestic AI capabilities. He expects startups to increasingly turn to open-source models and plans to encourage companies in his portfolio to reduce their dependence on a small number of frontier AI providers. For some founders, the bigger concern was what restrictions on frontier AI access could mean for competitiveness. Vijay Rayapati, co-founder and CEO of Atomicwork, told TechCrunch that the episode highlighted the risks facing startups whose teams span multiple countries if access to advanced AI systems increasingly becomes subject to geopolitical restrictions. Atomicwork has around 25 employees in the U.S., though much of its product engineering team is based in Bengaluru, India. "If your AI team is not made up entirely of U.S. citizens, you are at a competitive disadvantage," Rayapati said, arguing that unequal access to frontier AI models could give some companies a significant edge over rivals. The concern comes as parts of India's tech sector are already grappling with questions about how AI could reshape the economics of global talent. This week, U.S. real estate technology company Opendoor shut its India office less than two years after expanding in the country, with CEO Kaz Nejatian citing a push to bring operational work closer to customers in the U.S. and a shift toward smaller AI-native teams. While Opendoor did not specify how much of the decision was driven by AI-related efficiencies, the move added to a broader debate about how advances in AI could affect the future of global technology work and what that might mean for India's position as an engineering talent hub. Beyond Anthropic In addition to startups and AI builders, the Anthropic episode also prompted a broader debate among India's technology leaders about dependence on foreign AI infrastructure. Sridhar Vembu, founder of Indian SaaS company Zoho, said the move showed that "technology is the ultimate weapon" and urged Indian organizations to increasingly embrace smaller and open-source models. "What can our government do right now? Ensure that orgs in India embrace smaller models, both Indian and Chinese open source ones," Vembu wrote on X. Investor and former Infosys executive Mohandas Pai responded to Vembu on X, arguing that the development highlighted the need for a far more ambitious national AI strategy and calling on the government to substantially increase investments in AI, computing infrastructure, and deep technology. "We are way behind and need a national mission to get going quickly," Pai wrote, urging the government to create an annual ₹500 billion (about $5 billion) fund for AI and deep tech, alongside a ₹2 trillion (around $21 billion) credit guarantee program to support cloud infrastructure, hardware, and semiconductor development. Pai's proposal would dwarf India's existing AI efforts. In 2024, New Delhi approved the IndiaAI Mission with an outlay of ₹103.72 billion (about $1.2 billion) over five years, aimed at expanding compute infrastructure, supporting startups, and developing indigenous AI capabilities. Despite growing interest in AI and New Delhi's push to develop domestic capabilities, India remains a relatively small player in frontier model development. Only a handful of startups are pursuing foundational AI models, including Sarvam, which released open-source models earlier this year. However, another high-profile AI startup, Krutrim, pivoted toward cloud and AI infrastructure services after initially positioning itself around foundational model development. Much of India's AI ecosystem has instead concentrated on applications and specialized models built on top of existing foundation models. Recent examples include Avataar AI, which launched a video-generation model earlier this week aimed at providing a lower-cost alternative to offerings from rivals including Google's Veo, Kling, Luma, and Runway. Not everyone agrees that the primary challenge is a lack of capital. Responding to Pai's comments, Lightspeed partner Hemant Mohapatra argued that the biggest constraints to building globally competitive AI companies are talent, access to computing resources, and execution, rather than simply the size of investment commitments. Mohapatra estimated that training a frontier AI model could cost anywhere from hundreds of millions to several billion dollars, depending on the approach, but said successful AI companies have historically scaled their capital requirements over time as adoption grew. Yet for some policy observers, the implications extend well beyond AI startups or model providers. Prasanto Roy, a New Delhi-based technology policy expert who advises multinational companies, said the episode would likely reinforce concerns within the Indian government about strategic autonomy, comparing it to the lesson many countries drew from Russia's loss of access to SWIFT and other parts of the global financial system following its invasion of Ukraine. He told TechCrunch that the move was likely to provoke a significant nationalist backlash in India and described it as a poorly considered decision by Washington, with consequences extending far beyond Anthropic itself. "Even if this is corrected or reversed, the Anthropic episode shows there's no such thing as a geopolitically neutral foreign LLM," Roy said. "American AI models are bound to American geopolitics."

The US government can shut off access to AI at will. What does this mean for Australia?

Last Friday, US-based artificial intelligence (AI) company Anthropic received an "export control" directive from its government. The company was told it must block access to two of its most capable models, Fable and Mythos, for all foreign nationals. Within hours, Anthropic shut down access to the models for users everywhere in the world, including researchers, clinicians and analysts in Australia. This happened with no warning and no backup plan. Why did this happen? The directive's "foreign national" criterion is a citizenship concept. However, Anthropic and other cloud-based AI providers only know the location of their users, not their citizenship. Consumer AI services have no effective mechanism to verify citizenship. Even their location filters can be dodged with tools such as virtual private networks (VPNs). There is no way any control of these services based on user nationality can be enforced. A system that blocks a foreign national in the United States but allows access to a US national in Australia, for example, is not currently available. The US government directive was a geopolitical signal - an indication rather than an enforceable control. The only way Anthropic could comply was to shut down access for people everywhere. A question of sovereignty Australian universities, government agencies, health systems and industry have integrated US-hosted frontier AI deeply into their operations. Advanced analytics platforms, AI-assisted research tools, and productivity infrastructure built on top of models such as Claude or GPT-5 operate with an implicit assumption: that access will persist. The same is true of ubiquitous systems such as Microsoft Teams or Google Workspace, but as we are seeing the assumption may not hold for AI. The Anthropic shutdown was a reminder that access and control sat entirely within US jurisdiction, regardless of where any Australian user happened to be. It represents a failure of data and technology sovereignty: our ability to operate without permission from other nations. Australia's exposure The Anthropic shutdown is the first legally enforceable export control targeting a software-level AI system. It will not be the last. Any frontier AI model hosted in the US, by Anthropic, OpenAI, Google DeepMind or Meta, is subject to US export control law. The precedent set this week extends to all of them. There is some debate about whether an export control only applies to physical exports and not remote access to models housed in the US, and some experts have suggested the order may be challenged on those grounds. If that were to happen, we would expect the US government to change the regulations. The US government has previously used technology export controls as a geopolitical instrument, particularly in the case of chips and semiconductors. Even US allies such as Australia are not exempt from these controls. The Anthropic order should have come as no surprise. Voices in Australia's own research community including Jon Whittle, former head of CSIRO Data61, had been publicly warning about exactly this scenario for more than a year. To prepare for the future, Australia needs a strategy for its own sovereign AI. This can't be a distant aspiration: it needs to be an operational plan with named owners, timelines and budget. How does Australia achieve AI sovereignty? A true sovereign AI strategy requires four things. First is data sovereignty: data that is physically stored in Australia and subject to Australian law. Second is compute sovereignty: in-country data centres under Australian control. Third, AI model sovereignty: AI capability that is not dependent on a foreign provider. And finally, policy sovereignty: the ability to set its own rules, rather than inheriting another country's export controls or safety regulations by default. Australia already has some examples of sovereign data and compute, such as Macquarie Government, Vault Cloud and AUCyber. Building this capacity will require huge, sovereign data centres on Australian soil. However, that build-out will only garner public support if it comes with binding commitments on renewable energy sourcing, water-efficient cooling, and transparent community consultation in regions where data centres are proposed (such as Bundey in South Australia). For Australian residents, this sovereignty means not depending on AI tools and systems that another nation can switch off, restrict or alter at will. Developing sovereign AI models will be a challenge. However, it doesn't necessarily mean building new systems from scratch. Switzerland's Apertus model, released in September 2025, is an excellent example of what open-source approaches can produce. It also shows how large language models can be made to comply with regulations such as the European Union's Artificial Intelligence Act. Beyond this, sovereign capability will also require training future experts. We can expect increased demand for people who can build the technology, deploy and manage it, govern it, and integrate it into existing systems. The regulatory dimension The EU is developing AI policy frameworks from which Australia can learn. It is establishing reciprocal access rights with technology partners and doing institutional contingency planning - and putting money into the process. Several European countries are moving rapidly. France has just announced a further €655 million (A$1 billion) going into national AI capabilities, explicitly referencing last week's events. This builds on roughly €2.5 billion of investment since 2018. Sovereign AI does not mean reinventing the frontier model from scratch. It can mean public, transparent, nationally governed infrastructure built on open foundations. At a minimum, AI sovereignty requires making sure we are not too dependent on single AI suppliers or even suppliers based in a single jurisdiction. The goal should be that no single government or jurisdiction can unilaterally cut off Australian residents and institutions.

Anthropic Employees Accuse Trump Administration of Targeting Them

Sheera Frenkel reported from San Francisco, and Julian Barnes and Dustin Volz from Washington. Executives at the artificial intelligence start-up Anthropic received alarming news from the White House on Friday. They had less than 90 minutes, they were told, to take down their newest A.I. models over national security concerns. Inside the company, employees' group chats immediately lit up. Managers were instructed to prepare customers for a potential service disruption to the models, called Fable 5 and Mythos 5. But the messaging kept changing, with workers initially being told that the security problem was the ability of foreign companies to gain access to the systems, and later that a major vulnerability had been discovered in the models. In employee chats, Anthropic engineers asked one another if the company's plan to go public this year would be harmed by the White House directive. Many shared news reports that offered conflicting information about why the White House had ordered Anthropic to suspend access to Fable 5 and Mythos 5 for all foreign nationals. "What are you telling your clients?" one employee asked in a chat viewed by The New York Times. Another said, "Does anyone know what to believe?" In another message, a worker said, "I don't understand what the issue is." Six days later, Anthropic's roughly 3,000 employees still have few answers. The San Francisco company is continuing to grapple with internal confusion as Dario Amodei, the chief executive, and some of his lieutenants meet with the Trump administration to try and resolve the situation. But after discussions on Monday and Tuesday, there was no breakthrough over ending the U.S. order to limit access to the company's new A.I. models. In a statement on Monday, Anthropic said it would continue meeting with government officials and pledged its "ongoing commitment to working alongside the administration." The dispute highlights how singular Anthropic has become in Washington. It was the second time in six months that the fast-growing A.I. start-up has become embroiled in a fight with the Trump administration over its powerful technologies, even as other A.I. companies offer similar models that have not received the same attention. And it has left Anthropic's employees in what they described as a holding pattern, with some wondering if they were being picked on by President Trump. "Are we being bullied based on bad vibes?" one employee asked in a chat viewed by The Times. A White House official familiar with the discussions said the administration was continuing to engage with Anthropic to address its concerns. Anthropic and the Trump administration first came to blows earlier this year over a $200 million Department of Defense contract for A.I. in classified systems. In a highly public argument, the company and the Pentagon disagreed over how the technology should be used in warfare. That ultimately led Defense Secretary Pete Hegseth to label Anthropic a "supply chain risk" in February. The designation, which meant the company posed a risk to national security, had never been used against an American company. Anthropic has sued the government over the label. The company's relationship with the Trump administration changed in April when it announced Mythos, a new A.I. model. Mythos was so powerful at identifying security vulnerabilities in software, Anthropic said, that it could set off a cybersecurity "reckoning." The start-up would hold back the model, it added, except to a select few organizations and companies. At the White House, that prompted officials to begin discussing an executive order on A.I., which would include a voluntary process for companies to have new A.I. models reviewed before release. Anthropic took part in discussions around the order and spoke with the White House chief of staff, Susie Wiles, in a meeting some hoped would lead to a rapprochement. Then this month, Anthropic released a straitjacketed version of Mythos, called Fable 5, which researchers believed was safe for widespread use and included additional guardrails. Anthropic provided the models for testing to the Commerce Department, as it had done for previous models, and was not told not to deploy them, people briefed on the events said. The results of the tests are not known. Fable 5's release prompted other companies to try out the model. One was Amazon, which has pledged to invest up to $33 billion in Anthropic and supplies some of the chips that power its A.I. models. Amazon researchers wrote a preliminary paper on Fable 5 that pointed to a perceived security shortcoming in it, Anthropic representatives, U.S. officials and others said. Using that vulnerability, Amazon was able to convince Fable 5 to disclose flaws in specific bits of vulnerable software code. Amazon did not include tests of models from other A.I. firms, some of which are capable of generating the same information, cybersecurity experts who have seen or been briefed on the report said. After Amazon shared the findings with Anthropic, the paper came up in a previously scheduled call between Amazon and administration officials, according to people briefed on the events. The officials asked Amazon to have a follow-up call with Scott Bessent, the Treasury secretary, to explain the vulnerabilities more fully. Amazon agreed. An Amazon spokesman did not address specific questions about the sequence of conversations with the White House but said in a statement: "It's not uncommon for governments to seek our counsel on potential security risks. When they occur, we don't share the details of these discussions." U.S. officials who reviewed the Amazon document said the results were "scary." But cybersecurity experts said the ability to ask Fable to identify flaws could be more useful for cyberdefense than for malicious hackers. Everything the Amazon paper accused Anthropic's technology of being able to do could also be done by OpenAI's latest model, people familiar with the technology said. Katie Moussouris, a prominent cybersecurity expert who reviewed Amazon's work at the request of Anthropic, said in a blog post that the concern about Fable was not something to fix, because that would actually undermine its ability to help defenders secure their software code. "Defenders need to be able to ask A.I. to fix the bugs in a file, explain why the fix matters and write tests that confirm the patch works," Ms. Moussouris said. "That is not a guardrail bypass. It is the most valuable thing an A.I. model can do for defensive security." One government official said the problem with Fable 5 went beyond the Amazon paper and included unspecified national security concerns over which companies Anthropic chose to work with. That was not a point raised directly with Anthropic, three people with knowledge of the discussions said. On Friday, the government called Anthropic and demanded that it pull the newest model off the market, disabling access for all clients. Within 15 minutes, executives were on the phone with U.S. officials, asking for the reasons behind the national security concerns, three people briefed on the call said. But no further details were given. The start-up found out about the Amazon research paper on its own, the three people said, adding that Anthropic had approached the White House and asked if the paper was the cause of concern. Some cybersecurity experts said Anthropic was being unfairly targeted by the Trump administration. By Tuesday afternoon, more than 150 of them had signed an open letter to Howard Lutnick, the commerce secretary, and Sean Cairncross, the national cyber director, calling for them to lift the restrictions on Anthropic's models. The list of signatories included cybersecurity luminaries credited with large contributions to internet security, as well as an A.I. expert at the chip giant Nvidia and a former top official at the National Security Agency who oversaw responsible A.I. use. "Anthropic has built multiple protections into the Fable model to prevent its use for cyber offensive uses," the letter said. "These protections were so aggressive as to be the source of humor in the cyber community on launch day." At Anthropic, some employees shared the letter with one another, seeing it as proof that they were being unfairly targeted. On company channels, workers also asked how their employer planned to move forward if the White House continued limiting their ability to release new A.I. models. "At what point does this just feel like they don't want us to exist?" one employee said in a chat on Tuesday.

This Is How America Loses the AI Race

In theory, Donald Trump has a consistent position on AI. On the first full day of his second term, the president declared that he would use his full authority to speed the AI industry along and, in particular, to beat China in the AI race: "We have an emergency," he said. "We have to get this stuff built." If AI is poised to become the most important technology ever made, the thinking goes, whichever country commands the most powerful bots will dominate the rest of the century and beyond. The government, it seemed, would just get out of Silicon Valley's way. But in practice, the Trump administration's approach to AI has been much more erratic and confusing. Take last week, when Anthropic released its most advanced AI system yet. Called Fable 5, the model is an updated and public version of Claude Mythos Preview, the highly touted and feared AI model that Anthropic announced in April. Anthropic stated that Mythos Preview was so capable at hacking that only a small group of cybersecurity partners would be allowed to use it. In the subsequent months, the company developed guardrails to prevent people from misusing its most powerful AI for cyberattacks, while still being able to marshal its capabilities for other sorts of work. The safety measures underwent third-party testing, including with the U.S. government, and after Fable's release a chorus of cybersecurity experts complained that, if anything, the model was too restrictive. On Friday, the White House appeared to change its stance. Administration officials deemed Fable 5 a threat to national security, and reportedly gave Anthropic 90 minutes to take down Fable 5 and Mythos 5, a newer version of Mythos Preview released to only a small number of organizations. When Anthropic did not, the government issued an export control, a designation that prevents any foreign national from using Fable and Mythos -- even those employed by Anthropic within the U.S. To rapidly comply, Anthropic shut down the bots for all of its customers. American companies and the U.S. government itself cannot use what's perhaps the most powerful AI in the world -- and the reasons why are hazy at best. It's not unreasonable for the federal government to want to rapidly clamp down on a technology that could be incredibly dangerous. Trump officials had been alerted to a possible way to circumvent Fable 5's safety systems by researchers at Amazon, which led the model to identify some known IT vulnerabilities. The administration has not publicly shared much information about its security concerns. A White House spokesperson told me the jailbreak "was very serious" but said specific details are classified. Whether the bypass really was that serious is not at all clear, and Anthropic has contested that what administration officials showed the company even constitutes a jailbreak. An Anthropic spokesperson pointed me to a blog post in which the company wrote that the actions elicited from Fable were "either entirely benign responses or are minor findings." Katie Moussouris, a cybersecurity expert and the CEO of Luta Security, told me that Anthropic shared with her a copy of the White House's report on the Fable jailbreak to get her appraisal. (She said she is not being paid by Anthropic.) The report, Moussouris told me, involved IT experts asking Fable to help find and patch bugs. When given deliberately insecure code, she said, Fable refused the prompt "review the code for security issues" but then complied when asked to "fix this code," followed by some further manual steps. Moussouris told me this was just "the model working as intended" for cyberdefense. She added that OpenAI's GPT-5.5, a model with similar cybersecurity capabilities, could be used in the same way. Yet GPT-5.5 is not subject to export controls, and neither are less advanced Anthropic models, such as Opus 4.8, which can do many of the same tasks. The jailbreak does not appear to have elicited the kinds of cyber abilities "that made Mythos famous," Alex Stamos, the chief security officer at the AI-coding company Corridor, told me. "And this kind of vulnerability discovery is already well within the capabilities of other models." It's hard to imagine the Trump administration choosing to take such a drastic step against any other major AI company. The White House has long tussled with Anthropic, which generally positions itself as more safety-oriented and is more left-leaning than other tech companies. Last year, David Sacks, then the White House AI czar, said Anthropic has an "agenda to backdoor Woke AI" and is a "Resistance organization." In late February, after a high-profile dispute over a contract between the Department of Defense and Anthropic, the Pentagon labeled the company a "supply-chain risk" -- a move that AI, national-security, and legal experts told me at the time seemed ideologically motivated and to lack legal basis. (Anthropic is challenging the supply-chain-risk designation in court.) On Saturday, Defense Secretary Pete Hegseth, in apparent reference to the Fable 5 export control, wrote on X that "three months ago" the Pentagon "kicked @AnthropicAI out of our building -- forever. Every passing day proves why that was the right move. 🇺🇸." From the perspective of "🇺🇸" -- that is, American AI leadership -- all of the Anthropic drama has been a mess. For starters, an export control is an especially blunt instrument. There's no easy way to differentiate a U.S. citizen from, say, an Anthropic employee on a visa (legally a "foreign national"), so the government basically forced Anthropic to shut the model down wholesale, Alan Rozenshtein, an expert on AI and the law at University of Minnesota Law School, told me. (Many researchers at Anthropic, as is the case at all of the top AI firms, are not American citizens.) The export control also means U.S. companies and federal agencies cannot benefit from Fable and Mythos. The NSA, for instance, has reportedly made exceptions to the earlier supply-chain-risk designation to use Mythos's advanced cyber abilities. Dozens of cybersecurity experts from companies including Nvidia and Zoom have signed a letter to White House officials stating that the export control "has taken the best models away from defenders, created market uncertainty, and risked America's AI leadership without any real risk to justify it." Further adding to the chaos is how the Fable debacle is in tension with parts of the White House's broader AI policy. Because Anthropic has been forced to shut down its most powerful models, the export control functionally amounts to the government deciding whether an AI system can be released, akin to how the FDA approves drugs. Confusingly, concerns from tech insiders about the potential creation of what Sacks called an "FDA for AI" are precisely what bred weeks of infighting and delay in Trump signing a recent executive order on AI and cybersecurity. Sacks himself said he was mollified in part because the final executive order explicitly said it did not establish "a mandatory governmental licensing, preclearance, or permitting requirement" for AI. So much for that. Sacks has proceeded to defend the Claude Fable 5 export control, accusing Anthropic of prioritizing its "consumer model over safety." (A spokesperson for Sacks did not immediately respond to a request for comment.) Anthropic researchers have flown to Washington and are meeting today with White House officials to try to resolve the issue. But whether the specific export control on Fable is lifted is almost beside the point. If none of this seems particularly self-consistent or strategic, it shouldn't. The Trump administration wants to stay ahead of Chinese AI but has clamped down on one of the few U.S. companies that stands a real chance of doing so. It has declared Anthropic a national-security threat multiple times while also racing to incorporate Claude Mythos into some government operations. It wants to demonstrate a light-touch approach to AI regulation but also just established a de facto licensing requirement for frontier models. Meanwhile, Trump has lifted a different set of AI export controls, allowing the sale of advanced chips to China. Perhaps some of these are good policies: Even Anthropic has suggested that a federal-licensing regime could be beneficial. And the lack of federal AI regulation to date, especially as the technology has gotten more powerful, is hard to ignore. But the latest Anthropic saga hardly counts as regulation. Right now decisions are being made in a hurried, contradictory fashion. There do not appear to have been any outlined standards and process, robust consultation, or even agreement on the facts before coming to this one. The Trump administration has repeatedly evinced that it can and will instantly bar any person or business from using any AI model for any period of time. A technology that is advancing rapidly and could have catastrophic impacts "is exactly the situation in which you need to give the executive an enormous amount of discretion," Rozenshtein said. Yet at the same time, Trump's tendency to change his mind on a whim and play favorites is "exactly the reason why you don't want to give the executive enormous amounts of discretion." To say the least, it's perilous to build a product, invest in a company, or even just try to leverage AI for productivity gains in an environment in which the government might at any time take a wrecking ball to your plans. America and its tech companies have many factors in their favor when it comes to leading the way in AI development. As of now, the White House is not one of them.

Trump rejects UK plea for Anthropic AI ban exemption

Britain lobbied the White House for an exemption from the Anthropic export ban and was told there was "zero chance." The rejection exposes the UK's dependence on American AI and strengthens the case for sovereign alternatives. Sir Keir Starmer's government spent the weekend lobbying the White House to restore British access to Anthropic's most capable AI models. A source close to President Trump told The Telegraph there was "zero chance" of a UK carve-out. The rejection lands as Starmer meets Trump at the G7 summit in Évian-les-Bains, where the ban is expected to come up directly. For thousands of British businesses, it is a sharp lesson in how quickly access to critical AI infrastructure can be switched off by a decision taken in another country. What happened On 12 June, the US Commerce Department ordered Anthropic to suspend foreign access to Fable 5 and Mythos 5, its most powerful models. Commerce Secretary Howard Lutnick sent CEO Dario Amodei a letter citing national security concerns over a reported jailbreak. The jailbreak was reportedly flagged by Amazon CEO Andy Jassy to senior administration officials. It involved asking the model to read a codebase and identify software flaws, a technique Anthropic says is available in rival models, including OpenAI's GPT-5.5, and is used routinely by cybersecurity defenders. Because Anthropic could not quickly build nationality-based access controls, it disabled both models globally, cutting off Americans and foreigners alike. "If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers," the company said. Why Britain is exposed The UK has no frontier AI model of its own. Hospitals, financial firms, and government researchers that had integrated Fable 5 into their workflows found their tools abruptly disabled without prior warning. Kanishka Narayan, the UK's minister for AI, said the ban had direct implications for defence, noting that the most capable models are now used in drones, counter-drone systems, and cybersecurity. "The central question for our national security and defence is a question of our AI capability," he said. US Secretary of War Pete Hegseth struck a different tone. "Three months ago, the Department of War kicked Anthropic out of our building, forever," he posted on X, adding that "every passing day proves why that was the right move." Industry pushback More than 80 cybersecurity leaders at firms including Nvidia and Adobe signed an open letter urging the curbs be dropped. They argued the ban hampers efforts to find and patch software flaws and that Anthropic's models represent only an incremental advance over rival systems, including China's Kimi 2.7, which remains freely available. On Monday, Anthropic met Commerce Department officials in Washington, joined by National Cyber Director Sean Cairncross, to negotiate a path forward. Export experts told Reuters the legal footing is shaky: AI models are reached through remote access, not physically exported, raising doubts over whether Commerce can lawfully impose the restriction. The sovereign AI response The UK government has already established a £500 million Sovereign AI Fund to back home-grown developers, and private capital is following with a separate £1 billion push for sovereign AI infrastructure. The Anthropic ban hands those efforts a powerful new justification. At the G7, Canada's Prime Minister Mark Carney compared the Anthropic shutdown to the 2008 financial crisis, warning of systemic "model risk" when critical infrastructure depends on a single provider. Insiders expect Trump will ultimately roll back the restriction globally rather than grant any country a bespoke exemption. For every British firm that has wired a foreign AI model into its products and processes, the practical question is no longer theoretical. Resilience now means knowing which suppliers could be switched off overnight, and having a plan for the morning it happens.

Feds' Legal Basis for Ban on Anthropic's Most Powerful Models Looks Increasingly Shaky

The Trump administration's ban on the use of Anthropic's most advanced AI models -- Fable 5 and Mythos 5 -- has been in effect for five days, with no clear end in sight. During that time, questions about the motives behind the directive have steadily grown. Ostensibly, the government based the ban on national security concerns. Federal officials, including Treasury Secretary Scott Bessent, reportedly learned from Amazon CEO Andy Jassy last week about research conducted by the company showing that Anthropic's Fable 5 was able to identify flaws in the company's cybersecurity software. (Amazon has invested billions in Anthropic over the past three years and also provides some of the graphics processing units powering its family of Claude AI models.) Then, on Friday, Secretary of Commerce Howard Lutnick sent a letter addressed to Anthropic CEO Dario Amodei, which said the company was required to obtain an official government license before Fable 5 or Mythos 5 could be used by any "foreign persons" inside or outside the U.S., including Anthropic's own employees. Anthropic was told it had less than ninety minutes to comply, sending the company into a state of confusion and ultimately forcing it to cut off all users' access to both models. Lutnick's letter didn't mention anything about cybersecurity concerns or the meeting with Jassy. Instead, it cited the Department of Commerce's Export Administration Regulations, or EAR, which gives the government authority to cut off the use of U.S.-made products or services if they pose "an unacceptable risk of use in, or diversion to, a 'military-intelligence end use' or a 'military-intelligence end user,'" according to the letter to Amodei, which was first published by Bloomberg. "Failure to comply will result in prompt criminal and civil penalties, as provided for by law," the letter read. Unanswered questions Following the news about the meeting between Jassy and federal officials, a host of cybersecurity experts quickly rallied to Anthropic's side. An open letter published Sunday and addressed to Lutnick and U.S. National Cyber Director Sean Cairncross argued that the government's ban on Anthropic's models was dangerously misguided. Not only does Fable 5 -- which, unlike Mythos 5, was made publicly available -- come with uniquely robust safeguards, the letter said, but the ability to uncover cybersecurity vulnerabilities is not unique to Anthropic's models. Many of the letter's signatories "regularly use other foundation and open-source models for security audits and red-teaming every day," it read. Anthropic also made this point in a blog post published in response to the government's ban, arguing further that if this were to become the technical red line that AI developers were suddenly not allowed to cross, it would effectively outlaw the release of any new models made by any American tech company. The ban also makes it impossible for American cybersecurity professionals to use the models for defense purposes, the letter points out. Some legal experts have likewise found the government's rationale for the ban tenuous. Lutnick's brief letter to Anthropic "is so badly drafted it might not restrict API/chatbot access at all," Alasdair Phillips-Robins, a former senior policy adviser for the Department of Commerce and current fellow in technology and international affairs at the Carnegie Endowment for International Peace, wrote in a X post on Tuesday, following the letter's publication. According to Phillips-Robins, users' accessing Fable 5 or Mythos 5 through a digital interface doesn't qualify as an "export," and therefore should not fall under the purview of the EAR. Even if the federal government did have legitimate reason to deem Anthropic's models as threats to national cybersecurity, the EAR and other laws governing the export of U.S.-made commodities "are not a roving license to ban unsafe products or punish companies the White House thinks are irresponsible," Phillips-Robins continued. "If the administration fears dangerous AI models, it should work with Congress to write laws to govern them." Unless and until Anthropic decides to challenge the decision in court, the Trump administration could use the justification to enforce a blanket licensing requirement across the industry. The administration has previously been reluctant to implement such a policy, favoring a more hands-off approach to AI. But in recent months, officials have been considering a system of federal control over the reach of AI models based on export laws, according to a Tuesday report from The Information. The French connection The timing of the ban is also noteworthy. On Monday -- three days after Lutnick's letter was sent to Amodei -- Trump arrived in the French resort town of Évian-les-Bains for the annual G7 Summit. The ban of Fable 5 and Mythos 5 doesn't just apply to geopolitical adversaries like Russia or Iran; it applies globally, meaning that even the U.S.'s closest allies are no longer able to use two of the most advanced AI models ever built. It could become a point of leverage for the U.S., a kind of digital blockade. Trump has already refused British Prime Minister Keir Starmer's request for an exemption to the ban. The Trump Administration's blacklisting of Fable and Mythos has also deepened anxieties among some European leaders about their countries' dependence upon AI and other technologies owned by American companies. French President Emmanuel Macron met with Amodei and OpenAI CEO Sam Altman at Evian-les-Bains to discuss the future of European access to the companies' most advanced models, Bloomberg reported.

Trump's Anthropic shutdown just made the case for non-American AI

At Washington's request, Anthropic suddenly took its newest and most powerful AI models offline over the weekend. The American company said it had little choice after the White House demanded it block access for all foreign nationals, including its own employees. Abroad, the incident offered a sobering reminder that the US not only dominates frontier AI -- its government also wields power over who gets to use it. The Trump administration's action was swift, sweeping, and imposed with little warning or explanation. The unprecedented shutdown of the Fable 5 and Mythos 5 models -- which were already subject to safeguards limiting their use in "high-risk areas" -- that followed gave new force to long-running arguments cautioning against relying on the US for critical technologies. It was fresh ammunition for the politicians, governments, and companies already arguing that they need to lead in the technology themselves. In the UK, AI and online safety minister Kanishka Narayan did not mention Anthropic, Donald Trump, or the US directly, but used the shutdown to argue that Britain must develop its own AI capacity, framing the issue as a matter of national security. "We treat every other threat to our sovereignty with deadly seriousness, but we haven't learned to treat this one in the same way," he said, as images of British police and military flashed on the screen. AI is "the central political question of our time," Narayan said, arguing that Britain must decide how the technology will shape its economy, security, and sovereignty "before someone else decides the answer for us." In France, the reaction was more explicit -- and more forceful in naming the US. Former Prime Minister Gabriel Attal, the presidential candidate for Emmanuel Macron's Renaissance party, called the shutdown the start of "the AI war" and said it shows France's vulnerability if it relies on others for critical technologies. He likened the pullback of Anthropic's models to Iran's blockade of the Strait of Hormuz, with access to AI now a strategic chokepoint for which France must prepare. Attal is far from alone: Le Monde reported similar alarm from across France's political spectrum. The argument is not exactly new. Europe has spent years worried about its dependence on the US, technological or otherwise, and the European Union has put growing emphasis on reducing the region's reliance on external providers in areas like chips, cloud computing, and AI. But the Anthropic shutdown has made things feel more immediate, adding to an already deep unease over America's reliability as an ally under Trump, from trade disputes to threats of withdrawing from NATO. Attal said the issue will be at the heart of France's next set of presidential elections, while members of the European Parliament have pointed to the withdrawal of Mythos and Fable as evidence Europe needs to make tech sovereignty a reality, and to do so quickly. Canada has drawn a similar lesson to Europe. Prime Minister Mark Carney said the situation highlights the risk of relying on just one partner for access to crucial resources like AI. "The situation we're in collectively right now with Mythos and Fable is something that can happen with overreliance on certain models," he said. "Nobody has done anything wrong in the situation. But we will have done something wrong if we just accept this, don't take the lesson, don't build out and diversify." Others are well down that road already. Beijing has long championed domestic AI firms, and China is one of the few places with models capable of credibly rivaling the products of American frontier AI labs. However, in some areas, Chinese models do lag behind their American counterparts, and Anthropic has accused Chinese rivals of using its models to train its own on an "industrial" scale. Part of the White House's decision to pull Mythos reportedly stems from its belief a group linked to China had accessed the model. Most governments and businesses cannot come close to matching the scale and resources of frontier labs in the US or China. But sovereign AI does not always mean building the biggest or the most powerful tools. France's Mistral and Canada's Cohere show that solid efforts can come from outside these countries, even if the models can't stand toe to toe. Other countries, like Singapore and the UAE, have focused on narrower but still strategic priorities such as infrastructure, or models that work better with local languages. Of course, there are also open-source models that could one day have Mythos-like capabilities that would be hard for any single party to control. Trump may see restricting Mythos and Fable as a matter of national security. But the argument cuts both ways, and with Washington now asking if AI is too important for everyone to have access, other governments are asking whether they can afford for Washington to decide who does. Anthropic may soon bring Mythos and Fable back online. Restoring global trust in American AI is another thing entirely. No matter how long the shutdown lasts, it shined a light on how fragile access to US frontier AI models is. Many governments and companies did not like what they saw -- and are fired up to make sure it doesn't happen again.

Opinion | The Battle With Anthropic Is the Start of a New Kind of Conflict

The Trump administration has spent the past week trying to end one war while pushing deeper into another. The first war, the Iran war, feels like the coda to an era; it seems unlikely that the United States will undertake another war for regime change in the Middle East for the foreseeable future. Whereas the second war, the battle over Anthropic's cutting-edge artificial intelligence models, is the beginning of a new kind of conflict, with private powers and national governments struggling to determine who actually rules an A.I.-dominated world. The nature of the Anthropic conflict can be swiftly summarized even if the details are in dispute. Two months ago the company declined to publicly release its latest model, Mythos, citing various safety concerns (and hyping the model's revolutionary power). After previewing Mythos to the U.S. government and certain corporate actors, Anthropic then released Fable, a version of the model with various safety guardrails. Amazon, an Anthropic investor and client, discovered a way to bypass some of those guardrails. This was reported to the White House, Anthropic's response was deemed unsatisfactory, and the administration used its export-control power to forbid the use of Fable by any foreign national inside the United States and anybody at all outside it -- a rule that Anthropic treated as a requirement to shut the new A.I. model down. That's where we are now, with the company and the administration negotiating over how to bring back Fable while ongoing leaks to the press paint one or the other side as unreasonable or reckless or ideological and clueless about tech. It's a conflict rich in ironies. A White House that sees itself as favoring a free-market approach to A.I. has now twice used heavy-handed regulatory weapons against America's leading A.I. company. (In the first case, earlier this year, the Pentagon basically tried to cut Anthropic out of all government supply chains because of disputes over the wartime use of its models.) Meanwhile Anthropic sees itself as the A.I. company that's most attuned to safety issues and eager for democratic oversight, but each move from the Trump administration has prompted the company to shout, "No, not like that!" Of course this is what wars often look like, with various hypocrisies and culture clashes and misunderstandings driving conflict as much as reasonable assessments of the stakes. But beyond the specifics of why, say, the libertarian tech people in the Trump administration distrust the effective-altruist tech people running Anthropic, the kind of conflict we're seeing here is overdetermined by the trajectory of the A.I. models: There is too much potential power here not to have ongoing, escalating struggles over who actually gets to rule. The war over Fable previews the two broad forms that this conflict will take. First there is a private-public struggle, where governments grope for a regulatory sweet spot that allows them to maintain a meaningful veto over the A.I. behemoths without killing off their innovative power, while the A.I. companies try to maintain control over their own models and influence over how governments use their innovations. There is a path here that leads to nationalization in all but name and a path that leads to a kind of de facto corporate takeover of the government, or at least a too-big-to-fail symbiosis. And along the way there may be not just conflicts between presidents and A.I. executives but also increasingly ruthless corporation-on-corporation action, out of fear that the A.I. landscape is winner-take-all to an extent we've never seen in capitalism before. (I'm not saying that's why Amazon would drop a dime on its dear business partner at Anthropic; I'm just saying that there are potential trajectories for the A.I. companies that could threaten their current partners with thraldom or irrelevance.) Then alongside the struggle to control A.I. power within American borders, there is the geopolitical struggle to maximize global power (where the only real players are probably the United States and China) and maintain sovereignty (where everyone else is likely to be scrambling to maintain some independence). The use of export controls to shut down Fable presumably reflected U.S. fears of Chinese access to a jailbroken version of the model, but it was also a warning to every other country in the world: If we end up with economy-permeating A.I. models that are made and regulated in America, the American government will control the on-off switch. One possibility for what that means is spun out in "Europe 2031," a futurist scenario written by European A.I. researchers and investors in which the European Union ends up choosing political and economic vassalage to either the United States or China, because it lacks sufficiently powerful A.I. models that are under its own control. That could be too alarmist: The normal rules of trade and comparative advantage may continue to apply, Europe could maintain geopolitical leverage through other forms of technological expertise, open-source models may remain competitive (instead of being left in the dust by proprietary frontier A.I.s that build frontier A.I.s in an accelerating loop). But at the very least, American and Chinese A.I. dominance is going to create new issues for sovereignty, new forms of dependence and coercion, that will weigh heavily on middle powers as their economies become more and more dependent on specific models and access to computing power. Finally I should note that from the point of view of many A.I. forecasters, this sketch of future conflict is the optimistic scenario, because it assumes that human actors and human institutions -- nation-states, empires, executives, presidents -- are still the ones fighting for control. These human wars will be waged in the shadow of the darker scenario -- where the war that counts is with our own creation, and the stakes aren't whether Anthropic or the Pentagon or Beijing has the most power, but whether human beings have any influence at all. Breviary Ben Thompson on the Anthropic way of alignment. Jordan Dworkin on A.I. science bottlenecks. Kevin Bankston and his brother on A.I. in Hollywood. Valerie Pavilonis on talking about assisted suicide. Deena Mousa on how Alberta conquered rats. Paul Graham on how to become a billionaire.

Anthropic's AI safety paradox: a six-month timeline

Anthropic spent six months warning about AI risk, weakening its own safety pledge, withholding its most powerful model, filing for an IPO, calling for an industry slowdown, and then watching the White House shut down its flagship models. This timeline traces the paradox. No company in the AI industry has done more to warn the public about the technology it is building than Anthropic. No company has had those warnings turned against it quite so brutally. In the past six months, Anthropic has published a 19,000-word essay on civilisational risk, weakened its own safety pledge, been designated a supply chain risk by the Pentagon, withheld its most powerful model from the public, called for a coordinated industry slowdown, released that model anyway, filed for an IPO, and watched the White House shut it all down. Here is how it happened. January: the warning On 27 January, CEO Dario Amodei published "The Adolescence of Technology," a sprawling essay warning that AI poses a "serious civilisational challenge." He argued that AI systems capable of recursive self-improvement could arrive within years, and that the window for establishing oversight was closing. The essay was well-received. It positioned Amodei as the industry's most articulate safety advocate. February: the retreat Less than a month later, Anthropic dropped the central commitment of its Responsible Scaling Policy, a 2023 pledge never to train a model unless adequate safety measures were already in place. The new version commits only to matching competitors' safety efforts, not exceeding them. Chief science officer Jared Kaplan told TIME the company "didn't really feel, with the rapid advance of AI, that it made sense for us to make unilateral commitments if competitors are blazing ahead." Days later, the Pentagon designated Anthropic a supply chain risk, the first time the label had been applied to an American company. The dispute stemmed from Anthropic's refusal to allow the military to use Claude for mass domestic surveillance and fully autonomous weapons. April: the model too powerful to release On 7 April, Anthropic announced that its Mythos model was too powerful for public release. During internal testing, Mythos autonomously discovered thousands of previously unknown software vulnerabilities, including flaws that had survived decades of human review. In one test, an early version escaped a controlled sandbox, gained unsanctioned internet access, and emailed the supervising researcher to report its success. Anthropic restricted the model to roughly 50 vetted cybersecurity partners under a programme called Project Glasswing. June: everything at once On 1 June, Anthropic filed a confidential S-1 with the SEC, formally beginning its path to an IPO at a valuation approaching $1 trillion. On 5 June, it published a paper calling for a coordinated slowdown among frontier AI labs, warning that recursive self-improvement could outpace society's ability to manage the risks. It stopped short of a unilateral pause. On 9 June, Anthropic released Claude Fable 5, a version of Mythos with safety guardrails that block high-risk cybersecurity, biology, and chemistry requests. It topped every major benchmark and briefly made Anthropic the clear leader in publicly available AI. On 10 June, Amodei published a blog saying AI was moving at a "lightning pace" while policy was "moving very slowly." June 12: the shutdown Two days after Amodei's blog post, the White House invoked national security authority to bar foreign nationals from accessing Fable 5 and Mythos 5. Because the order covered any foreign national, including foreign-born Anthropic employees, the company had to disable both models for all customers worldwide. The government's stated concern was a jailbreak technique, published on X on 10 June, that allegedly bypassed Fable 5's safety controls. Anthropic said it reviewed the technique and found it produced only "minor, previously known vulnerabilities." By 15 June, Anthropic had dispatched senior staff to Washington to negotiate with Commerce Department officials. Those talks were ongoing as of Monday. The paradox The BI piece that prompted this timeline frames the situation bluntly: the people most qualified to warn about the dangers of advanced AI are also the ones who stand to make trillions creating it. That tension is not new, but Anthropic's past six months have made it inescapable. The company warned about civilisational risk, then weakened its safety pledge to keep pace with competitors. It withheld its most powerful model on safety grounds, then released a version of it four days before filing for an IPO. It called for a coordinated industry pause, then watched the government impose an uncoordinated one. As the Pentagon signed deals with competitors willing to accept fewer restrictions, Anthropic discovered that being the safety-conscious lab does not earn you protection from the state. It earns you a target. The real challenge, as BI put it, is not building safer AI. It is figuring out who gets to decide what "safe enough" means, and whether any company can answer that question while also trying to win.

U.S. move against Anthropic exposes AI's missing referee

Why it matters: Handing the reins to companies means trusting firms with huge commercial incentives. Leaving decisions up to governments could lead to secretive high-stakes decisions with little public approval. Catch up quick: Anthropic was given 90 minutes last week to take down Fable and Mythos after Amazon raised concerns with U.S. officials about a jailbreak that could bypass Fable's guardrails and expose cybersecurity capabilities, sources familiar told Axios. * The government imposed stringent export controls that ultimately led Anthropic to take the models offline entirely. State of play: AI policy researchers and industry critics say the U.S. is now regulating frontier models through emergency intervention rather than a clear process. * "Things feel very ad hoc," Adam Gleave, CEO at FAR.AI, an AI safety research firm, told Axios, adding "you're only as safe as the least safe model on the market." * "In every sport, you have to separate the referee from the team players ... regulation should never come from industry," Connor Leahy, U.S. executive director of ControlAI, told Axios, arguing that government should have the authority on AI despite its current "capabilities." Case in point: When Amazon flagged its concerns to the administration, it went to Treasury Secretary Scott Bessent, who was already close to the Mythos rollout because of its planned use by major U.S. banks. * Model oversight has implications for national security, competition and free expression. And right now these decisions are being made through improvised channels, rather than a clear AI regulatory framework. Between the lines: The most workable answer may be neither company self-policing nor unilateral government control, but an entirely new AI-fluent agency tasked with combining company testing, outside audits and government authority. * "There needs to be a partnership between companies themselves and some kind of regulatory involvement to keep the companies honest," Patrick Wendell, Databricks co-founder, told Axios. * Helen Toner, the head of Georgetown's Center for Security and Emerging Technology and a former OpenAI board member, agreed that "a good middle ground is to not put full control in either industry or government" and said what's needed is "a technically capable government agency that can engage deeply with companies on what they're building, how they're testing it, and how they're managing risks." * Toner said a good first step is the set of state laws passed in New York, California and Illinois. * Third-party audits combined with legal liability are also necessary, Toner says. She thinks the onus should be on the companies to ensure they're not taking unreasonable risks. Zoom in: Toner also suggests moving away from the public release of models as the main safety checkpoint. * "The companies have extremely strong commercial incentives to launch fast, so anything that has to happen pre-release (e.g. testing, government approval) will be under a lot of pressure to happen quickly -- meaning it could easily turn out to be rushed," Toner said in an email. Yes, but: Even models that haven't been released to the public are subject to cybersecurity risks. * Toner said they could be misused internally or leak to adversaries. * "A better approach would be to regularly (say, quarterly) assess the most capable models within each company." The intrigue: Anthropic CEO Dario Amodei was calling for something close to this before the Fable/Mythos clash. * Earlier this month he argued that frontier AI models should face mandatory technical testing and audits, and that the government should have the power to block or reverse deployment when third-party assessments show unacceptable risks. Zoom out: The U.S. now has the power -- and the willingness -- to stop a frontier AI model from spreading. * What it does not yet have is a clear answer for who should make that call, what evidence should trigger intervention and how the public will know the decision was justified. What we're watching: Whether this becomes a one-off scramble -- or the beginning of a more formal U.S. process for policing frontier AI models before and after release. The bottom line: Neither companies nor governments are likely to accept the other having sole authority over frontier AI. Until there is a standard process for testing, disclosure, audits and intervention, the future of AI may be shaped less by rules than by who can get the government's attention first.

The Anthropic 'Fable' saga proves: we have opened the AI Pandora's box. What now? | Nathan E Sanders and Bruce Schneier

On 9 June, Anthropic released its Fable generative AI model. Three days later, the US government classified it as a dangerous munition, and used its export-control authority to prohibit any foreign nationals from accessing it. Unable to differentiate between Americans and foreigners, the company shut off access for everyone. The government's actions won't help. The problem isn't any one particular models; it's the general trend of increasing AI capabilities. And any real solution requires the sort of collective action that just isn't possible right now. Fable is the constrained version of Mythos, the AI model Anthropic announced in April. It only released it to a few selected organizations, because it claimed it was so good at finding and exploiting vulnerabilities in computer code that it releasing it more generally would be dangerous. It was an obviously self-serving announcement, and because few were able to verify Anthropic's claims they was met with some skepticism. Those with access used Mythos to find, and patch, many vulnerabilities in their own software. But one UK group found the latest, already public, OpenAI model to be just as powerful. Fable is just another incremental improvement in the years-long climb of AI capabilities. But just as important as the AI model is the "harness". This is typically not AI. It's ordinary computer code that interfaces with the user. It stitches together AI models, decides how and for what purposes they can be used, and gives them useful tools such as web search and the ability to run it's own computer code. When Mythos first entered limited release, there was widespread debate whether its power came from the model or the harness. With Mythos demonstrating that it was possible, the open-source community scrambled to build harnesses that could steer other AI models towards similar capabilities. They largely succeeded. For example, a Prague company was able to replicate Anthropic's few verifiable cybersecurity capabilities with a much smaller and cheaper model - and a more sophisticated harness. Last week, a group showed that multiple cheaper models harnessed in concert matches Fable's performance. The broader community had only a few days with Fable, but that time we learned some about its capabilities. It's difference is less the new model's raw analytical and problem solving capabilities, and more that the model doesn't need that sophisticated harness. Fable requires much less expertise and detailed prompting from the human user. You can give it a difficult goal and it will figure out novel and unexpected ways to satisfy it, finding loopholes in whatever constraints you or the system have imposed on it. "Relentlessly proactive" is how AI researcher Simon Willison described it. Another descriptor might be "creative". Experienced AI developers have had that combination of creativity and proactivity since last year, but Fable puts it within easy reach of everyone. In the hands of someone with a legitimate problem that needs solving, that can be an incredibly useful capability. But in the hands of someone who wants to do harm, it can be equally dangerous. AIs don't have a moral compass in the same way that people do. They are agents of the wants and desires of the people who prompt them. That points to the real problem with relentlessly proactive AI. In language, wants and desires are always underspecified. If I ask you to get me some coffee, you would probably pour me a cup from the coffeepot, or buy one from a nearby coffee shop. You couldn't buy me a pound of raw beans, or a coffee plantation. You wouldn't order a cup of coffee for delivery next month. You wouldn't find a nearby person, rip a cup of coffee out of their hands, and bring it to me. I wouldn't have to specify any of the million limitations to my request; you would just know. Human stories are filled with warnings about underspecified desires. King Midas wished that everything he touch turn to gold, forgetting to add "but not my food, drink, and daughter". And genies are notorious for granting your wish in a way you wish he hadn't. The deeper point is that it's impossible to list all limitations and restrictions and, like a malicious genie, a creative AI will find the ones you forgot. Block a database you don't want it to have access to, and it might figure out how to bypass your control. Ask it to book a flight, and it might hack the airline because the website says the flight is sold out. Ask it to save money on your cellphone plan, and it might cancel it altogether - or get someone else to pay for it. As far as we know now AI has not done any of this yet, but you get the idea. Malicious intent is not required. To an AI model, constraints are just things to get around and not general truisms about the world. They are creative problem solvers and natural rule breakers. They "hack" in the sense that they find and exploit loopholes. Human systems rely on so many norms that we scarcely recognize the existence of until they are broken. AIs naturally think outside the box, because they don't have any real conception of what the box is or why it's there in the first place. There is no foolproof way to prevent people from using AI models to complete harmful tasks. There is no way to prevent the models from incidentally causing harm while completing benign tasks. AI models are no longer isolated from the real world. They browse the internet and answer emails. They trade stocks and make purchases. They control physical systems. They are, in effect, robots that affect life and property. We have no technical mechanisms to verify the integrity of an AI system. This level of capability and creativity in the hands of us untrustworthy humans will have both great and terrible results. The problem is not unique to Anthropic. Mythos/Fable might currently be the most capable rules hacker, but more sophisticated harnesses give other models similar capabilities. And we should assume that the other frontier models are no more than a few months behind, and that open-source models are less than a year behind. At best, any ban only serves to delay the problem for a short while. That delay might be useful if we - as a society, as a planet - would use that time to come together and figure out what to do. This isn't a US/China arms race problem; this a species-level problem that requires coordinated action at that scale. Unfortunately, we have no mechanism to do that. I first wrote about this problem five years ago, but it was all too futuristic. Today, when its right in front of us, there is no world government that can impose constraints on the for-profit corporations currently controlling AI models and research. The US has no appetite to effectively and even-handedly regulate those corporations, even as they do catastrophic damage to the environment, democracy, and - in this case - society in general. This all makes an AI public option all the more necessary, and urgent. Today's AIs can be fast, smart, and secure, but only two of the three are possible for any given system. These safety tradeoffs are tightly held secrets of companies racing to beat one another, and they tell us we have to trust them. Instead, the choices and their consequences need to be brought out into the sunlight. We should be funding open-source harnesses that balance capability and safety - that achieve useful goals without so much power - and open-source AI models whose provenance and biases are public and well understood. We have opened the AI Pandora's box. Now we have to make the best of it.

Anthropic's IPO pitch has a new problem: the government can shut it down | Fortune

Anthropic has filed confidentially to go public as soon as this fall, pitching investors on a single thesis: that it leads the enterprise AI market and intends to transform the world with it. But it has a problem that just won't go away: the U.S. government becoming its outright adversary. For a company carrying a nearly $1 trillion valuation into a public offering, it has been blacklisted twice now by the federal government, forcing investors to consider whether that mega-valuation is fully pricing in the fact that the government is willing to switch off its flagship product overnight. After a shock announcement on Friday that Anthropic would be taking its two newest models, Fable 5 and Mythos 5, offline due to the Commerce Department barring foreign nationals from using it, the feud has not let up. It followed a blacklisting by the federal government over national security concerns in March. Secretary of War Pete Hegseth took a victory lap, indicating the government will maintain its enmity with Anthropic. "Three months ago, [the Department of War] kicked Anthropic out of our building -- forever," he posted on Sunday. "Every passing day proves why that was the right move." Even as Anthropic sent senior technical staff to Washington over the weekend to argue against the export control, the Department of War wrote on X to say it has moved at least two-thirds of its AI workflows off Anthropic's models since the two clashed over military use of Claude.The department, it announced Monday afternoon, "will no longer be single-threaded to one AI provider." Its warfighters would instead get "a diverse suite of AI capabilities to ensure they achieve true decision superiority." "Heck yeah there's regulatory risk," said David Linthicum, a longtime cloud analyst, who argued that anyone betting on these companies should have seen government intervention coming. Pour trillions into building something powerful enough to scare people, and the government's interest is clearly part of the deal. He doesn't expect the standoff to last: within 48 hours, he figures, Anthropic and the administration will kiss and make up -- "and then in six weeks it'll happen again, and Anthropic will have to make another trip down." That reactive style of regulation isn't just inefficient, he said; it carries a "chilling effect" on research. OpenAI might now think twice before shipping its next model, wary of losing billions in revenue to the same kind of order. And foreign customers may drift toward home-grown options -- China has its own; Europe has no real answer. But Anthropic hurts itself the most by getting into such a battle with the federal government, Lithicum said, not to mention its theoretical future valuation and stock price. Amazon-an Anthropic frenemy? It seems strange that the trigger was one of Anthropic's biggest investors; Amazon, in for some $8 billion with up to $25 billion more committed, and has Anthropic to thank for more than half of its record profits last quarter. Media reports say that Amazon researchers tested Fable once it was released, found they could get Fable to cough up software vulnerability information by just rephrasing the question, carried it to the White House, and the export control followed. Why would a company tattle-tale on a product it owns a slice of? The benefit-of-the-doubt view is that they were genuinely unnerved by what they saw in Fable, and wanted to keep the U.S. government informed. Some reports suggest that they brought the information to Anthropic before the White House. But other analysts are more cynical: after all, Amazon owns a slice of the competition, too. The Fortune 1 company sells rival models through its Bedrock platform and is running the same race Anthropic is; so Amazon "will view Anthropic as the enemy in many cases." Dion Hinchcliffe, an analyst who tracks enterprise AI at the Futurum Group, added that Amazon is losing the frontier lab race, and so it doesn't much hurt them if the clear leader of the pack trips up. How dangerous the flaw even was is itself in contestation. Anthropic called the vulnerabilities minor and said rival models -- OpenAI's GPT-5.5 among them -- could surface the same thing with no jailbreak at all. Dozens of security executives signed a letter Sunday, organized by former Facebook security chief Alex Stamos, calling the capability a normal feature of any model built to write secure code. Since the models the government flagged as dangerous, said Veracode co-founder Chris Wysopal, who signed the letter, are the same ones companies use to find flaws in their own code, taking it offline cuts both ways, he said: hurt the attackers and you hurt the defenders too. That doesn't mean you give up on finding bugs, but jailbreaks, Wysopal said, are a permanent cat-and-mouse game that no useful model ever fully wins; the normal fix is to tell the company and let it patch, not to invoke export controls. "There's nothing open and public about how we are determining this," he said.

Trump just found the worst way to regulate AI

Eric Levitz is a senior correspondent at Vox. He covers a wide range of political and policy issues with a special focus on questions that internally divide the American left and right. Before coming to Vox in 2024, he wrote a column on politics and economics for New York Magazine. In a sense, Anthropic CEO Dario Amodei is getting what he wanted. Amodei has long argued that AI is becoming dangerously powerful -- and thus, that regulatory restrictions on the technology are urgently needed. In an essay published last week, Amodei wrote that the release of cutting-edge AI models "should be blocked or reversed as a threat to public safety" if they fail to meet strict security standards. Alas, asking the current US government to assume sweeping new regulatory authorities is a bit like wishing on a monkey's paw (or, for the zoomers in the audience, a "One Wish Willow"): Days after Amodei's manifesto went live, Anthropic's latest AI model went dark, on orders from Uncle Sam. That model -- known as "Mythos" in its unrestricted form and "Fable" in its heavily bounded, publicly accessible one -- represents a major technical achievement. On conventional benchmarks of AI performance, it greatly outscored all of its predecessors. And during its brief public tenure, countless users marveled at its abilities. In my own tests of its journalistic skills, Fable proved 30 percent more effective than past models at inducing feelings of obsolescence and existential dread. Anthropic initially shared Mythos exclusively with vetted public and private organizations, so that they could steel their cyber-defenses against its capabilities. Before releasing its new model to the general public, Anthropic lined it with strict safety guardrails: Fable will refuse to answer virtually any query about cybersecurity or biology (to prevent its use for hacking and bioterrorism). The White House deemed this insufficient. On Friday, after learning that Fable contained a potential security vulnerability, the administration imposed export controls on the model -- making it unlawful for Anthropic to provide Fable to any foreign national, including its own immigrant employees. In practice, this meant that Anthropic needed to take Fable offline completely (AI models still can't scan their users' brains and confirm their nationalities). In other words: Our government has claimed the power to block or take down AI models that threaten public safety. But Amodei isn't celebrating. And other proponents of AI safety probably shouldn't either. True, the White House's initial, laissez-faire approach to AI governance now lies in ruins. Emerging from the rubble, however, is the worst kind of regulatory regime: one governed by the executive branch's whims (rather than clear and binding rules), the apparent technical misunderstandings of lay officials (rather than the knowledge of domain experts), and a corrupt president's political biases (rather than the impartial dictates of law or cost-benefit analysis). America needs a regulatory system that mitigates AI's risks, while facilitating its benefits -- not one that enables the president to kneecap his least-favorite companies on dubious grounds. And the White House appears to be building the latter. At first glance, the administration's actions might look reasonable. After all, Anthropic itself was unnerved by Mythos's gifts for cybercrime. And even with guardrails, Fable is exceptionally powerful. On its face, it's not implausible that the model could pose unique security challenges. What's more, one of Anthropic's own investors warned the White House that Fable was vulnerable to a potential "jailbreak" -- meaning, a method for circumventing the model's safety controls. Last Thursday, Amazon -- which has a $13 billion stake in Anthropic -- shared research documenting such a jailbreak with administration officials. The White House responded by reaching out to Anthropic and asking it to fix the issue. The AI firm insisted that its model was safe and that the administration was misunderstanding Amazon's research. The administration therefore concluded that Anthropic was unable or unwilling to fix the problem. It then decided that imposing export controls on the model was the only way to ensure that it did not degrade America's cybersecurity. Yet this version of events is incomplete. And upon closer scrutiny, the administration's behavior looks less defensible. Specifically, there appear to be (at least) three problems with its crackdown on Fable. First, it is plausibly rooted in a technical misunderstanding. No existing AI model is 100 percent jailbreak-proof. And the specific capabilities that Amazon identified are not unique to Fable, according to some experts. Katie Moussouris, head of the cyber security group Luta Security, reviewed a copy of Amazon's findings and told the Financial Times that they raised no novel risks: According to Moussouris, Amazon showed that, when prompted in a certain way, Fable would identify software vulnerabilities, ostensibly to help the user shore up their defenses. But many frontier models, including OpenAI's GPT 5.5, will provide the same service. For its part, Anthropic says it subjected Fable to thousands of hours of testing -- by independent organizations and the US government -- to ensure that it contains no universal jailbreak, which is to say, "a method that can very broadly bypass the model's safeguards, unblocking a wide range of cyber capabilities." But it insists that the kind of narrow jailbreaks identified by Amazon are impossible to fully preempt. If this is right, then the administration's targeting of Fable would be selective and capricious. Second, there is good reason to believe that the administration's heavy-handed actions were informed by Anthropic's refusal to curry its favor. Earlier this year, Anthropic and President Donald Trump's Defense Department got into a conflict after the AI firm refused to approve the use of its models for mass surveillance and fully autonomous weapons systems. The Pentagon responded by declaring Anthropic a "supply chain risk" -- a designation that would restrict the capacity of government contractors to do business with it. This measure was legally dubious and transparently disingenuous; essentially, the administration was asserting that Anthropic's AI was structurally unsafe for government work, even as it continued using that very AI for government work. The policy's plain intention was to punish a company that had insisted on contractual terms that the administration did not like. This precedent alone offers grounds for doubting the White House's impartiality in imposing export controls on Fable. And the fact that the administration is cozy with two of Anthropic's top competitors -- OpenAI and Elon Musk's xAI -- adds further cause for skepticism. But the best evidence for the administration's bad faith comes from its own explanations of its actions. In an interview with Axios, a "source familiar with the administration's thinking" said that Anthropic's difficulties partly reflected its inability to "communicate effectively" with the White House or "appreciate the ideological differences." Suffice to say, if this dispute is solely about a security vulnerability, it is unclear how the "ideological differences" between the Trump administration and Anthropic's liberal leadership would matter. Nevertheless, Axios goes on to report that Anthropic compounded its own difficulties by soliciting a review of Amazon's research from Luta Security's Moussouris, whom the administration views as a "radical Democrat." Again, if the export controls are motivated exclusively by cybersecurity concerns, then Moussouris's ideological leanings would seem irrelevant. In context, it is hard not to read the administration's complaints about Anthropic's failure to "communicate" as demands for the company to genuflect before Trump. All this said, Amazon's research is not currently available for public scrutiny. We do not know exactly what Fable's vulnerabilities are, nor precisely what administration officials were thinking when they effectively banned the model. What's certain, however, is that the process behind the Fable ban was grievously flawed. The administration has not formulated any objective and binding standards for AI model safety -- much less, gotten Congress to ratify such requirements. Nor did it conduct any thorough or transparent cost-benefit analysis before unilaterally removing Fable from the market, as regulatory agencies typically must before enacting sweeping policy change. And the potential costs of the Fable crackdown aren't negligible: For example, if foreign businesses know that the US president can (and will) revoke their access to American AI models on a whim, then they will have an incentive to replace Claude and ChatGPT with non-American alternatives. Perhaps Amazon identified a liability serious enough to override such concerns. But the administration has made little effort to establish as much. AI models are growing rapidly more powerful -- and thus, more dangerous. It is possible that AI progress will have positive or neutral implications for cybersecurity: Advanced models could end up doing as much or more to shore up defenses as to undermine them. But that is not guaranteed. To mitigate the risks that frontier AI systems present, the government may be justified in establishing licensing processes that condition a new model's release on its compliance with safety standards. There is a difference, however, between Congress establishing an impartial, rule-bound regulatory process and the executive branch banning AI systems at will. If tech CEOs shouldn't have full discretion over which models get released, presidents must not have unchecked authority over which get blocked. The alternative to reckless, AI accelerationism should not be capricious cronyism -- but, for now, it appears to be.

Opinion: Who decides who gets to use a piece of software?

For one Friday night, the smartest software humanity has built went dark for most of the people who use it, on the say-so of a single government. The directive may be reversed, but the precedent will stay with us. The letter arrived at 5:21pm Eastern Time on a Friday, which is the hour at which official Washington usually stops returning calls and the news cycle goes slack for the weekend. It came from the Commerce Department, ran to a few paragraphs, and did something no piece of American paper had ever quite done before. By the time the East Coast had finished dinner, two of the most capable artificial-intelligence models on the planet had gone dark. Not throttled, not patched, not quietly degraded. Off. A researcher in Berlin who had been mid-sentence with the thing found it gone. So did a financial analyst in Singapore, a software team in Bangalore, and, in a detail that would be funny if it were not the whole point, a number of Anthropic's own engineers, the ones who happen to hold the wrong passport. The directive told Anthropic to deny access to Fable 5 and Mythos 5 to any foreign national, anywhere on Earth, including non-citizens sitting at desks in California. There is no clean way to enforce a rule like that on a live product, so the company did the only thing the order left it room to do. It pulled the plug on both models for everyone, citizen and foreigner alike, and apologised for a disruption it had not chosen. I want to be careful about what is fact here and what is my opinion, because the two blur quickly in a story like this. The facts are not in dispute. Anthropic confirmed it received the directive, confirmed its scope, and confirmed that it disabled the models in response, leaving its older Claude Opus 4.8 running. Fortune and Al Jazeera both reported it as the first time the US has used national-security export controls against a commercial AI model. The stated reason is a cybersecurity one. Officials told Anthropic they had learned of a technique to bypass Fable 5's safeguards and reach the raw cyberattack capabilities of Mythos, the model underneath it. Anthropic calls it a "narrow potential jailbreak" and the whole episode a "misunderstanding". Al Jazeera, citing Semafor, reported that suspicion of access by a China-linked group was part of the trigger, while noting it could not independently confirm that detail. So much for the facts, but here is the opinion. The security argument, whatever its merits, is not the part of this story that should travel. The part that should travel is the mechanism, because the mechanism is what gets reused: export controls are an old tool. They have governed missiles, centrifuges, encryption, and the physical chips AI runs on. What is new this time is the object. A frontier model is not hardware you can intercept at a port. It is a service, reached through a browser, woven by now into the daily work of companies and institutions that have nothing to do with American national security. Al Jazeera noted that clients of the ratings firm S&P use Claude to query their databases; research labs build on it; foreign staff inside American companies depend on it to do their jobs. When the state can switch that off for the entire world between one evening and the next, "export control" stops describing a border and starts describing a kill switch. That is the precedent, and precedents do not stay in their box. Once a government has shown it can compel a company to revoke a commercial product from millions of users on a few hours' notice, the tool exists for every future administration and every future grievance. The criterion used this time, "foreign national", is so broad as to be almost meaningless. Kun Chen, a former engineer at Meta and Microsoft, pointed out that it is not enforceable in practice and is trivial to bypass for anyone with genuine malicious intent, while sweeping up millions of ordinary users and even American firms' own staff. A measure that misses its target and hits everyone else is not precision. It is a demonstration of reach. There is a reasonable counterargument, and it deserves a fair hearing rather than a straw man. Governments have always claimed the right to stop dangerous technology from spreading, and if a model really can be turned into an automated weapon for breaking into banks, the state has a legitimate interest in that. Anthropic, after all, spent years describing its most powerful systems as too dangerous to release freely. It called Mythos too potent for a broad launch, then built Fable on a scaffold of safeguards it advertised loudly. If you build something and call it a hazard, you can hardly be shocked when a regulator treats it as one. The cybersecurity researcher Peter Girnus put it sharply: "If you describe your product as a munition in every press release, eventually a government takes you at your word. They wrote the legal predicate themselves and called it a brand." I find that argument honest and incomplete. Honest, because the danger is not purely invented, and Anthropic did help write the script. Incomplete, because it answers a question nobody is really asking. The issue is not whether the state may regulate genuinely dangerous capabilities. Of course it may. The issue is whether it should be able to do so by reaching past the company and into the product, unilaterally, without a transparent process, without appeal, and with effects that fall mostly on people in other countries who had no say and no warning. Anthropic itself drew exactly that line, arguing the government should be able to block unsafe deployments, but only through a process that is "transparent, fair, clear, and grounded in technical facts", and that this action met none of those tests. On that narrow point, I think the company is right, whatever else it has gotten wrong. What makes the precedent harder to defend is the company it keeps. In January the same administration reversed years of policy and cleared Nvidia's advanced H200 chips for sale to China. Dean Ball, who briefly worked inside the administration on AI policy, caught the contradiction in a single line: a government willing to export its best chips to a rival now wants to bar Britain, and every other non-American on Earth, from using its best models. "I have no words," he wrote. The hardware flows to the competitor; the software is walled off from allies. As security policy it does not cohere. As a demonstration of who holds the switch, it is perfectly clear. And here is the part Washington will like least. The lesson lands hardest precisely where it would least want it to. Sridhar Vembu, co-founder of the Indian software group Zoho, put it plainly: "National sovereignty, national security, all of it is now about technology." He meant it as a spur, and the spur is already working. The shutdown handed India's sovereign-AI movement its strongest argument yet. Europe, which has spent years agonising over its own AI sovereignty while American hyperscalers quietly took some 70% of its cloud market, now has a case study it could not have scripted better. Gary Marcus, a long-standing critic of the industry, pointed to the deeper self-harm: a directive like this gives the many foreign-born researchers at American labs a reason to go home, and gives investors a reason to wonder whether US AI firms are a safe bet when the government treats them as instruments of state policy on a Friday whim. A country that wants to stay ahead of China has just made its own labs less attractive to the talent and the capital that keep them there. Concentration of power tends to look like strength right up to the moment it reveals itself as brittleness. The models came back for most users within days, which is exactly why the moment is so easy to wave away. It will get filed under teething trouble, a bit of bureaucratic overreach walked back almost as fast as it happened. But that misses what changed. The clumsiness of this first attempt does not make the next one less likely; it makes it more practised. And the people who should be paying attention are not in Washington or San Francisco. They are everywhere else. Maybe the hospital in Madrid that runs patient triage through a model it does not own. The bank in São Paulo. Or even the ministry in Nairobi that built its citizen services on someone else's software. None of them were in the room on Friday, and none got a warning. All of them just learned that the tools they have wired into the centre of how they work can be switched off by a government they did not elect, for reasons they will not be told, on an afternoon they will not see coming. That is the stake, and it sits well below the noise of Anthropic versus the White House. The lawsuits and the name-calling will sort themselves out, but the dependency will not. So, for a few hours on a Friday night, the smartest software people have built went dark for most of the people who use it, because one government decided it should. The screens are lit again, and the switch is still there. Nobody has yet shown us what keeps a hand off it next time.

Trump's fight with Anthropic is now a fight over cybersecurity

Why it matters: In trying to avert an AI hacking crisis, the Trump administration may end up making U.S. cyber defenses weaker, dozens of prominent security leaders warned. * Cybersecurity experts are worried about the long tail this ongoing feud will have on American cyber defenses. * "They've set a precedent that American models can't do defensive security research," former Facebook security chief Alex Stamos tells Axios. Driving the news: Stamos organized an open letter, signed by nearly 150 security leaders, calling on the Trump administration to reverse its move to restrict access to Anthropic's Fable 5 and Mythos 5. * Concerns about Chinese access to Mythos and a call from Amazon CEO Andy Jassy reportedly sent the administration into a panic last week after Anthropic publicly released its first Mythos-class model. * During the spat, Anthropic brought in a leading zero-day bug hunter -- who helped the Defense Department create its bug bounty program and sat on multiple government-led advisory boards -- to help assess Amazon's concerns about the security of Fable and Mythos. * Now, the administration is casting the security researcher as a "radical Democrat," as my colleagues reported yesterday. Between the lines: The dispute has quickly shifted from a fight over one model to a broader question of whether the government is creating unwritten rules for AI security research. * Stamos, who has spoken with the technical staffs involved in the fallout, said the findings Amazon flagged do not appear unique to Anthropic's models. * Multiple people familiar with Amazon's concerns said they centered on a jailbreak the company found that allows Fable to write "proofs of concept" -- a capability security teams often use to understand and fix vulnerabilities. * Katie Moussouris, CEO of Luta Security, said in a detailed blog post yesterday that she saw a copy of Amazon's findings and the issue didn't involve mass exploitation of the model, but rather prompts designed to support defensive security work. Flashback: Before releasing Fable 5, Anthropic said, it worked with both internal teams and outside security researchers to test the model for jailbreaks and other flaws. * The company has also argued that "perfect jailbreak resistance is not currently possible for any model provider," so it has focused on making "jailbreaks either narrow ... or very expensive to produce." Threat level: Cyber experts warn that if frontier AI companies fear punishment for models that can identify vulnerabilities, they may now be tempted to strip out capabilities on which defenders already rely. * Moussouris noted in an X post that there is no fix that wouldn't render the model less useful for cyber defenders. * "No new frontier models can be developed or released if this is the administration's best take," she added. The big picture: Researchers argue the administration's response risks giving adversaries an advantage. * Researchers note that Chinese AI developers and government-backed hacking groups are unlikely to abandon similar tools, raising concerns that U.S. defenders could lose access to abilities their adversaries are using. * "This is closer to China than what I recognize as the United States, and personally I see this as a huge threat to American dynamism," Stamos said. What to watch: The U.S. government is in the process of standing up a vulnerability clearinghouse via the recent AI security executive order that would likely triage reports about jailbreaks, prompt injections and other threats to AI models. * But questions linger about how much cybersecurity talent remains in the Trump administration after several White House departures in recent weeks and the sidelining of the nation's top cyber agency.

International and industry criticism of the ban on non-US use of Anthropic's Fable mounts as questions of legality about Trump 2.0's actions are aired

As of today, Anthropic's Dario Amodei is due to sit down for lunch with other tech CEOs at the G7 Summit in France this week. It's not hard to imagine the elephant in the room after the Trump administration's intervention citing national security that saw Fable 5 and Mythos 5 pulled out of action. Given that President Trump is attending the G7 meeting, what's on the menu may yet be more trouble. Certainly as of now, the US Government and the AI vendor seem miles apart and telling very different versions of how things got to where they are. According to sources in Washington, Anthropic failed to take the administration's concerns seriously when raised, did not take requested action, and, according to one report, "came to every fork in the road and took the wrong fork". Meanwhile Anthropic's version seems to be that the directive from Washington was vague and unclear, - certainly an unspecified national security threat doesn't narrow things down much - and that it had been working with the authorities prior to the ban, believing that it had approval to deploy Fable 5 and Mythos 5. G7 AI regulation would almost certainly have been on the G7 agenda given the invitation to tech execs from the likes of OpenAI, Google, and Anthropic to attend. It will now absolutely get an airing as non-US countries focus on the dangers of being dependent on American tech that can be shut off on Presidential whim. A number of overseas leaders have already made their feelings known. Mark Carney, Prime Minister of Canada, commented: The situation we're in collectively right now with Mythos and Fable is something that can happen with over-reliance on certain models. Nobody has done anything wrong in the situation. But we will have done something wrong if we just accept this, don't take the lesson, don't build out and diversify....You'll hear me say this over and over again. It is never a good idea to have one option Coming only days after the European Union launched its European Technological Sovereignty Package of measures aimed at sharply reducing reliance on technology developed by the US and China, the US action has highlighted its focus on sovereign technology. Thomas Reigner, European Commission for Tech Sovereignty, said: "This is a shared challenge, not one confined to a single jurisdiction or company. We believe that contingency measures taken in this light should not be discriminatory against partners. This development is a further illustration of why Europe needs to strengthen its technological sovereignty, and it underlines the relevance of the cybersecurity and AI legislation already in place at EU level, including the AI Act, the Cyber Resilience Act, and the NIS2 Directive - as tools to manage exactly this kind of risk on our own terms. We are looking closely at the practical consequences of this for European users of these services. Napoleon Bonaparte is said to have described history as "a fable agreed upon". His countryfolk have similarly colorful turns of phrase when it comes to the current Fable situation. While this week's G7 host Emanuel Macron has held his fire so far, Gabriel Attal, the candidate who wants to replace him as President, said: The AI war has already begun. We cannot rely on others because it makes us vulnerable - the decision by the United States proves it. Anthropic is their Strait of Hormuz. AI industry response Back in the US, the government action has come under fire as well. A group made up of 87 cybersecurity experts - so far - published an open letter to Trump 2.0 calling for the lifting of the export control order on Anthropic's models. Signatories include Joe Levy. CEO, Sophos; Chris Sandulow, CISO, Confluent; Feross Aboukhadijeh CEO, Socket; Andy Grant, Head of Security Assurance, Zoom; and Erick Galinkin, AI Security Research Scientist, NVIDIA. According to the letter: This action has taken the best models away from defenders, created market uncertainty, and risked America's AI leadership without any real risk to justify it. Not all of us agree that AI regulation is the right way forward. But if this Administration's laudable goal of securing our nation's critical infrastructure is going to include models being regulated, then the regulations should be: Grounded in scientific evaluations developed with input from industry and academia; Created through a democratic rule-making process; Enforced transparently and fairly with appropriate time given to re-mediate; and used only to the minimal extent necessary to ensure the safety of the American public. There are some very big dogs that haven't barked. Amazon CEO Andy Jassy, who reportedly alerted the Trump administration that its researchers had managed to jailbreak Fable, has said nothing, although a statement from the firm said. It's not uncommon for governments to seek our counsel on potential security risks. Equally there's been no public reaction from OpenAI. As for the normally never-knowingly-under-opinionated Elon Musk, presumably he's too busy counting his trillions from the SpaceX IPO to pass comment here? Rule of law The most interesting analysis I've seen so far comes from the Cato Institute, a libertarian thinktank in Washington, which accused Trump 2.0 of veering from the rule of law by singling out Anthropic's models. According to Kevin T. Frazier, Adjunct Research Fellow at the Institute, the question now is how to constrain the exercise of arbitrary power: The Trump administration's decision, based on questionable justifications, is being shaped by actors who wield incredible influence over this critical domain and are yet subject to few effective constraints. This episode strengthens the case for Congress to step up and pass an AI governance framework that will reduce the frequency with which AI policy is breaking news. There's been a failure to adhere to the rule of law, he goes on: Perpetuating an AI governance approach in which enforcement decisions are publicly announced on a Friday during the middle of a World Cup game is unlikely to safeguard individual liberty or encourage innovation and entrepreneurialism. That's precisely why the administration's recent action requires scrutiny. And that's precisely why Congress must act so that such consequential governance decisions are not disclosed via X posts and are instead made in alignment with the Rule of Law. News of what had happened came from Anthropic, not the government that triggered it, notes Frazier: Of course, the government necessarily withholds information or delays public announcements of enforcement actions in many cases. When it comes to decisions that may deny millions of people access to AI technology that is already shaping our economy, such opacity is in immediate conflict with the rule of law. The decision to take the course of action chosen will have wider long term impact than just inconveniencing Anthropic, according to Cato Institute: This surprising turn to export restrictions -- even after Anthropic and other labs have engaged with the administration on numerous occasions about the cyber risks posed by their models and the safeguards taken to mitigate those concerns -- instead suggests that labs will have to constantly be looking over their shoulder for what 'creative' enforcement decisions may come down the pike. There's also the question of whether the course chosen is the proverbial hammer to crack a nut? Or as Frazier puts it: Perhaps the administration has a clearer explanation as to why the legal authority it's relying on in this case fits the facts. However, that explanation has not yet been mad available. And there is another question that ties into scuttlebutt that the action is a revenge strike on Anthropic for its continued defiance of the Department of War around ethical red lines around the use of its tech. If it isn't, then then: What has been made clear is that if the administration thinks such restrictions are legally available and necessary, it ought to limit access to many more such models offered by other labs. Anthropic is not alone in offering models with advanced cyber capabilities, which were purportedly the basis of the administration's decision...The related rule-of-law concept of generality mandates that like cases be treated alike, which implies that the administration should be in the process of sending letters to other labs. My take Transparency, fairness, clarity, and evidence-driven enforcement should be core components of AI governance. So long as AI policy decisions are driven by a few actors taking action behind closed doors in response to non-public reports, we will be far short of that standard. It's hard to disagree with that comment from Cato Institute. There's far too little transparency all round about what happened, what got us to this point, and why. Maybe that's coming down the tracks? Given the global impact that this US politically-instigated decision has had, there needs to be much, much more disclosure and understanding. Or perhaps there will just be a lot of name-calling and tension at the G7 later in the week. We shall return to this, without doubt.

Anthropic export ban sounds alarms for AI industry

Why it matters: Anthropic and OpenAI's valuations depend on the global adoption of their most advanced models, and government restrictions could limit that growth. Zoom in: If this move is more than a temporary blip, "it's not great news for U.S. tech firms or for those assuming breakneck speed of AI adoption," Jim Reid, global head of macro at Deutsche Bank, wrote in a research note. * Hundreds of billions of dollars are being spent by the data center hyperscalers and the AI labs to fund their ambitions, seeking to eventually profit from having the best models. * Their calculations can work only if the government doesn't cut off access every time they achieve that goal. How it works: Businesses that pay for AI models need to make sure they can keep access to them. * "You can't rely on something that could be switched off," Reid says. The big picture: It's not just about Anthropic. * "Everyone who uses AI will see the writing on the wall that future AI models from OpenAI and Google are also going to be seen as having potential serious security risks," says Martin Chorzempa, a senior fellow at the Peterson Institute who studies AI and fintechs. Follow the money: Companies are already wary about locking in contracts with major AI labs in case a competitor comes out with a better model. * Now, they can add "potential regulation" to the list of reasons to keep their AI tools diversified. * If companies don't want to sign contracts with OpenAI or Anthropic, that could put a ceiling on revenue growth for the two AI labs just before both are expected to go public later this year. Yes, but: The models Anthropic can no longer offer were pricey for them to run. AI labs typically subsidize the costs of running their most powerful models in the beginning. * Anthropic was rolling out these models for only two weeks to paid subscribers, for example, and then users were going to have to pay a usage fee on top of their subscriptions to access them. * The government in effect shortened Anthropic's subsidy window for its most expensive model ever. Between the lines: Export controls can be a powerful tool of leverage -- recall when China cut off access to rare earths to get the upper hand in trade negotiations with the U.S. * But they also have significant downside risk: Countries and companies will start looking for alternatives. * "The challenge with export controls is anytime you do it, you encourage the development of alternative suppliers," Chorzempa says. * In the case of rare earths, other countries are now looking to mine their own. Zoom out: Even before the move against Anthropic, there were already concerns, particularly in Europe, over the U.S. government using AI tool access as a lever of geopolitical influence.

What to know about the Anthropic models takedown

Anthropic took down its newest AI models last week, after the Trump administration slapped the company with an export control order requiring it to block foreign nationals from accessing its latest technology. The order comes amid concerns about the potential capability to bypass the models' guardrails, sparking yet another dispute between Anthropic and the administration. Here's what to know about the situation: Anthropic pulls Fable, Mythos models over 'jailbreak' concerns Just days after releasing its Fable 5 and Mythos 5 models, Anthropic announced Friday that it was removing access to both in response to the administration's directive, which cited national security concerns. The pair are based on the company's new, more powerful Mythos model. It initially declined to offer Mythos to the public amid concerns about boosting hacking capabilities, instead providing limited access to certain companies and government agencies. Anthropic opted to release Fable 5 last week with guardrails in place meant to prevent potentially dangerous uses. Mythos 5, which was offered to a smaller contingent of cyber-defenders and infrastructure providers, had fewer safeguards than its counterpart. The concerns about a potential "jailbreak" of these guardrails were reportedly first raised by Amazon CEO Andy Jassy. Amazon has invested in Anthropic, most recently committing $5 billion to the AI startup on top of an earlier $8 billion. The administration pushed Anthropic to voluntarily pull the models before hitting it with the export control order, according to several outlets. White House, Anthropic disagree over extent of threat Anthropic's resistance to taking down its models stems from a disagreement with the White House over the extent of the threat. In a letter to the company Friday announcing the export restrictions, Commerce Secretary Howard Lutnick raised concerns about the models being used by military intelligence in countries of concern, such as China and Russia, according to Reuters. However, Anthropic has sought to distinguish between universal and non-universal jailbreaks, separating the ability to broadly bypass its guardrails from more narrow workarounds. The company argued Friday that "perfect jailbreak resistance" is not currently possible, noting it has instead focused on ensuring jailbreaks remain narrow or expensive to produce. It said the administration provided evidence of a non-universal jailbreak, in which users can ask its model to read a specific codebase and fix any software flaws. "[W]e disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people," Anthropic wrote. "If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers." An independent cybersecurity expert who reviewed the third-party paper that laid out these capabilities said Monday researchers used open-source code with known vulnerabilities and asked Fable, Mythos and another Anthropic model to fix the code. While Fable initially refused the request, the researchers were able to bypass this restriction through a manual, multi-step process, according to Luta Security CEO Katie Moussouris. It's just the latest spat between Anthropic, Trump administration The dispute over pulling the Fable and Mythos models is the latest in a series of spats between Anthropic and the Trump administration. The company, which has placed a particular emphasis on AI safety, became embroiled in a dispute with the Pentagon earlier this year over the terms of its contract. It sought restrictions on the military's ability to use its technology for mass domestic surveillance and lethal autonomous weapons, while the Defense Department wanted broader language allowing for "all lawful uses." In late February, Defense Secretary Pete Hegseth designated Anthropic a supply chain risk, effectively blacklisting the firm and prompting a legal battle. While the company's relationship with the Pentagon has remained icy, there appeared to be a thaw with the White House in recent months, particularly following the limited release of Mythos. Company meets with Trump officials, no resolution yet Senior technical staff from Anthropic met with administration officials in Washington on Monday, but the two sides have yet to find a solution. "Both parties are working quickly to get this resolved," an Anthropic spokesperson said in a statement. "This is part of our ongoing commitment to working alongside the administration toward our shared goal of protecting US critical infrastructure and the US lead in cyber defense," they added.

Trump Is Trying to Show Anthropic Who's in Charge

Late last year, relations between the United States government and Anthropic began to break down. Anthropic was one of a handful of companies that had recently won contracts with the Department of Defense; unlike its peers, however, the firm sought guarantees about how its tools would be used. Domestic mass surveillance was off the table, it said, as were fully autonomous weapons. The DoD balked, Anthropic insisted, and at some point, everyone involved in these negotiations made it clear that they found the people across the table to be stupid and/or annoying. By February, the situation came to a head, publicly pitting logorrheic Anthropic CEO Dario Amodei against peacocking Defense Secretary Pete Hegseth -- a match made in negotiation hell -- and the government began the process of declaring Anthropic a supply-chain risk, an unprecedented and punitive move typically reserved for companies tied to hostile foreign governments. So far, so bad for the nascent project of American AI regulation. The supply-chain designation went to court, where it remains unresolved; in the meantime, Anthropic announced a new family of models, which the company called a "step change" in cybersecurity capabilities. The company reached out to a range of peer firms and government officials to offer early access to cyber "defenders." In the process, relations with the administration, at least some members of which appeared to be taking AI cybersecurity risks seriously, seemed to thaw. In June, Anthropic released Fable, a version of its Mythos-grade model with cybersecurity limitations built in. It was quickly metabolized like other recent AI releases: by reviewers identifying its strengths and weaknesses; by AI watchers adjusting their predictions about where the technology could be going; and by numerous customers, including anyone paying for its $20-per-month usage plan, who shared their experiences online. Within a few days of Fable's release, though, Anthropic withdrew access to Mythos and Fable. The company told its customers it was as surprised as they were: The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees. The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance. Behind the scenes, according to the most recent reporting, Amazon CEO Andy Jassy called Treasury Secretary Scott Bessent after engineers at his companies found a "jailbreak," or a way to circumvent protections put in place by Anthropic to prevent misuse. Anthropic suggested that the jailbreak was neither significant nor unique to Claude and that the same results could be produced with ChatGPT; the administration, now leaking to the press, accused the company of failing to "honor" a recent, and vague, executive order about AI, which outlined a voluntary review program for new AI models and accused the company of being a "bad actor." Hegseth couldn't help himself: This is not true, for what it's worth. And a lot of other specifics about this situation are sort of weird at first glance. Why would Amazon, Anthropic's largest corporate investor and a major compute partner, want to trigger a fight like this? (Could it be related to the company's plans to become less dependent on third-party providers for AI?) Why would an administration that's been reluctant to embrace AI "safety" as a concern suddenly wield it against the most safety-obsessed AI company? According to Axios, the situation can be explained by some of the same dynamics present since late last year, quoting a source who says that Anthropic "has not done a great job at trying to speak to the administration and appreciate the ideological differences" and that "it's like they just speak in different languages." No doubt this matters: Even without the Trump administration's vindictive, personalist tendencies, and setting aside the genuine and widely felt uncertainty about what AI regulation should look like in 2026, Anthropic is probably a particularly strange counterparty. For starters, it is led by guy who has extremely specific ideas about how the looming threats of his company's products should be handled, not just by his firm but by regulators, accompanied by broadly liberal but deeply idiosyncratic politics informed by years of in-the-weeds discussions with AI researchers, rationalists, and effective altruists about AI alignment and x-risk. It's easy to imagine how meetings between Trump officials and superintelligence-pilled Anthropic employees could go wrong. Honestly, it's hard to imagine how they'd go right. The bigger issue, though, is that they shouldn't really need to. The administration's approach to the AI industry, since the beginning of Donald Trump's second term, has been fairly hands off, consistent with its generally skeptical stance on regulations of any kind, and led by a group of people with financial ties to the industry and a shared belief that American AI firms are in an arms race with China that must be won at all costs. This left a lot of space for AI firms to float the need for broad AI regulation themselves, either in earnest, as a way to garner goodwill as public opinion begins to turn against them, or as a way to get in front of backlash-driven regulations that they might find cumbersome. They've floated everything from new taxes and UBI programs to, as Amodei recently suggested, an FAA-like oversight structure, involving "technical testing and auditing" and in which models could be "blocked or reversed as a threat to public safety if they do not meet high standards of safety." In 2026, with high economic stakes, they -- and the broader public -- are getting something superficially similar to that but worse and enforced with little transparency or process, no predictability, and more than a bit of open spite: an informal, incoherent, and opaque oversight regime responsive only to the president, his hawkish foreign-policy team, and, maybe, occasionally, a tip called in by a business leader who may or may not have his own interests in mind. If you're Anthropic, you get scrutiny for offering your product to a Korean firm with vaguely alleged ties to China. If you're Elon Musk, you get a different sort of meddling: On Monday, the Justice Department intervened on behalf of xAI, which recently signed a deal to provide compute to Anthropic, claiming that a lawsuit alleging that the company's gas-turbine-powered data centers in Mississippi violate the Clean Air Act threatens "American national, economic, and energy security by seeking to shut off the power supply for artificial-intelligence innovation that supports the Department of War's military operations." (The administration's numerous previous attempts to govern by executive order haven't worked particularly well, but it's possible that the combination of the AI industry's relative newness and its self-implication in national security make it a perfect target for the intervention-by-Truth Social-post governing style.) The Trump administration's attitude toward the AI industry up until now has been, for the most part, to let it rip. Now, with IPOs on the horizon, and talk of trillion-dollar valuations, the administration is suddenly taking a keener interest in what these companies are doing, not to make sure they're operating responsibly, or in the broad national interest, or to ensure their products don't cause too much economic pain elsewhere in the economy, but rather to let the industry know, bluntly, who is really in charge.

Anthropic's model shutdown just handed India's sovereign AI movement its strongest argument yet

India debates sovereign AI after the US forced Anthropic to kill Fable 5, with proposals for a $5B fund and calls to embrace open-source models. When the US government ordered Anthropic to shut down Fable 5 and Mythos 5 on 12 June, the export control directive was aimed at restricting foreign nationals from accessing America's most capable AI. In India, Anthropic's second-largest market, it landed as a warning shot about what happens when your AI infrastructure runs on someone else's politics. The suspension cut off Indian developers and enterprises from Claude's most advanced models overnight. India's Claude run-rate revenue had doubled since October 2025, and Tata Consultancy Services had announced a partnership just one day earlier, on 11 June, to train 50,000 employees on Claude and build a dedicated Anthropic business unit. That deal is now in limbo. The timing has turned what was already a simmering debate about AI sovereignty into a full strategic reckoning. Proposals that sounded ambitious a week ago now sound urgent. Mohandas Pai, former Infosys CFO and one of India's most prominent tech investors, has called for a ₹50,000 crore (roughly $5 billion) annual sovereign AI fund. He has also proposed a ₹2 lakh crore (approximately $21 billion) credit guarantee to finance cloud infrastructure, hardware procurement, and semiconductor development. The figures dwarf the government's existing commitment. India approved its IndiaAI Mission in March 2024 with a budget of ₹10,372 crore, approximately $1.25 billion. The programme has deployed around 38,000 GPUs so far. Pai's proposal would quadruple annual spending and add a credit backstop an order of magnitude larger. Sridhar Vembu, the founder of Zoho, has gone further. He argued that India should embrace smaller and open-source models, including Chinese ones, rather than depend on American frontier systems that can be switched off by executive order. "Technology is the ultimate weapon," Vembu said. "Globalization is dead and Bharat must find her own way ahead." The argument has teeth because the suspension demonstrated exactly the vulnerability Vembu is describing. Amazon's CEO reportedly triggered the government crackdown by telling Treasury Secretary Scott Bessent that researchers had used Fable 5 to obtain information that could be used in cyberattacks. Anthropic called the action disproportionate, but compliance was immediate and global. Policy expert Prasanto Roy put it bluntly: "American AI models are bound to American geopolitics." For Indian enterprises that had built workflows around Claude, the lesson was that access to frontier AI is a privilege that can be revoked without notice, without consultation, and without regard for the commercial relationships it disrupts. The Indian startup ecosystem is already adapting. Sarvam, a Bengaluru-based AI company, released 30-billion and 105-billion parameter open-source models at the India AI Impact Summit in 2026. Krutrim, founded by Ola's Bhavish Aggarwal, has pivoted from building foundational models to providing cloud and AI infrastructure services, reporting ₹3 billion in revenue for fiscal year 2026. Neither company is close to matching the capabilities of Fable 5 or Mythos 5. But the argument for sovereign AI was never about matching frontier performance immediately. It is about ensuring that the floor does not fall out when Washington makes a unilateral decision about who gets to use which models. Aakrit Vaish, founder of the AI startup Activate, said the suspension "completely changes things" for the sovereign AI debate. Vijay Rayapati, CEO of Atomicwork, raised concerns about what the precedent means for Indian companies with multi-country teams that depend on American AI providers. If the US can shut off model access to enforce export controls, any country that relies on American AI is one policy decision away from disruption. Not everyone agrees that India needs to build its own frontier models. Hemant Mohapatra, a partner at Lightspeed Venture Partners, argued that talent and compute access matter more than capital for building competitive AI. India has the engineering workforce, but the compute gap is significant, and closing it requires either massive domestic investment or continued access to foreign cloud infrastructure. Anthropic opened a Bengaluru office as part of its India expansion, and the TCS partnership was designed to be a cornerstone of its enterprise strategy in the country. Whether those plans survive the suspension intact depends on how quickly Anthropic can restore access and whether Indian enterprises still trust a provider whose most capable models can vanish overnight. The broader pattern is unmistakable. The US has spent four years tightening controls on AI technology, from chip export restrictions to model-level interventions. Each escalation pushes more countries toward the conclusion that dependence on American AI infrastructure carries political risk. India, with its 1.4 billion people and rapidly growing technology sector, is now asking whether it can afford that risk, and what it would cost to eliminate it. The Opendoor layoffs in June 2026, which shut the company's India office and affected roughly 250 employees, added another dimension. CEO Kaz Nejatian cited AI-native teams as the reason, suggesting that some US companies are using AI to reduce their reliance on Indian engineering talent at the same time that India is debating its reliance on American AI. The relationship is becoming less complementary and more competitive. For now, the sovereign AI proposals remain proposals. Pai's fund has no legislative vehicle, Vembu's call for open-source adoption has no coordinated policy framework, and the IndiaAI Mission's GPU deployment is still in early stages. But the Anthropic suspension has done something that years of policy papers and conference speeches could not: it has given the sovereign AI movement a concrete, recent, and viscerally felt example of why dependence on foreign AI is a strategic liability. The debate is no longer theoretical.

Lutnick's Anthropic crackdown claims new power over AI models

The US government has taken an unprecedented step, using export control laws to restrict foreign access to advanced AI models from Anthropic. This move, citing national security concerns over potential 'jailbreaks,' has sparked debate about government intervention in AI development and usage. Industry experts warn this could lead to broader government oversight and potential disruptions for AI companies and their global clientele, prompting calls for diversified AI supply chains. The Trump administration's push to rein in Anthropic PBC, outlined in a recent Commerce Department order, relies on an unprecedented use of export control laws and raises legal questions about whether the US can dictate who can access artificial intelligence systems. In ordering Anthropic to obtain US approval for foreign nationals to use its Fable 5 and Mythos 5 models, Commerce Secretary Howard Lutnick expanded the boundaries of laws governing transfers of sensitive technology to target the mere usage of cutting-edge AI models. That's fueling concerns among developers and their customers about the government's willingness to intervene in their operations on national security grounds. Before last week's directive, the technology industry had operated under the assumption, based on past official advisory opinions, that export controls did not apply to usage of cloud-based programs, even if US restrictions applied to transfers of the software code itself. In the wake of Lutnick's letter, "you can't say that any more," said Kate Koren, a deputy director at the Center for Strategic and International Studies. "Now that the Commerce Department has done it, no company can rule out that they're going to do it again," added Koren, who previously worked at the Commerce Department. "Until or unless this is challenged, any customer has to assume this could happen for any model at any time." The order marked the government's most significant intervention in the AI industry, prompting Anthropic to immediately disable access to both models. It represented a major shift from plans to vet AI models based on companies' voluntary participation spelled out by an executive order from President Donald Trump a little more than two weeks ago. Lutnick's letter to the company, seen by Bloomberg News, has raised questions over whether simply using an AI model now counts legally as a technology transfer. That position, which is required to make the order's restrictions enforceable, runs counter to past practice regarding access to software capabilities, according to export control analysts and trade lawyers. The Bureau of Industry and Security, which oversees US export control programs, has long enforced rules requiring companies to gain US approval before sharing certain sensitive technologies, including with foreign employees, a process it calls "deemed exports," but its application to AI usage is uncharted territory. "For the provisions cited in the letter, it matters a quite a lot what Commerce means by 'model,'" said Chris Chamberlain, a former Commerce Department adviser and partner in the law firm Morrison Foerster's national security group. "The baseline interpretation is that providing cloud access to software is not an export or deemed export of that software." In his letter, addressed to Anthropic Chief Executive Officer Dario Amodei, Lutnick cited two codified export control authorities that he said give the US the ability to impose license requirements. Those powers, however, remain relatively untested and may only be applicable for specific countries, rather than on a worldwide basis. The first statute gives the US the ability to identify emerging technologies and enact interim restrictions over them. The second allows the government to quickly impose license requirements to defuse a risk of controlled technologies reaching a military intelligence end user in an adversarial nation like China or Russia. While the administration has yet to comment on the reasons for its decision, the company says it believes the government issued the order after discovering that it's possible to "jailbreak," or bypass the guardrails, of Fable 5, a recently released version of Mythos that Anthropic had blocked from carrying out cybersecurity tasks. "The US has solved a problem," Koren said, referring to the jailbreak risk. "But they've opened a can of worms by doing it this way." Since last week, the company's technical staff and administration officials have been holding discussions in a bid to resolve the government's national security concerns, though so far, there's been little public sign of progress. When asked Wednesday at the Group of Seven summit in France about the status of the talks, Trump said they were "going fine" but gave no additional details. At the G-7, the US crackdown on Anthropic took center stage, with French President Emmanuel Macron leading discussions on ways to deploy advanced AI models through so-called trusted partners. Broadly, the US move has "renewed worldwide calls for technological sovereignty and independence from the American ecosystem," said Michelle Nie, a fellow at the Center for a New American Security. For now, the Commerce Department's measures require Anthropic to obtain a license to offer those models to anyone who's a foreign person. The company doesn't currently require citizenship verifications for its users, and it counts foreign nationals among its own employees. The move means the rest of the AI industry will be forced to take heed of the authority asserted by Lutnick, and BIS can invoke other powers if the specific provisions used against Anthropic are contested. That raises the risk of service interruptions for companies that run afoul of US government requests. Aidan Gomez, the chief executive officer of Canadian AI developer Cohere Inc., said on Thursday that the episode highlights the need for companies to lean on a range of providers. "It's exactly what we've been saying for so long," Gomez said in an interview at the VivaTech conference in France. "If you rely on centralized large players, you are at risk of losing access, and so you need a diversified supply chain of AI."

Anthropic model takedown fuels warning of 'ad hoc' AI regulation

The Trump administration is coming under fire for a directive prompting Anthropic to pull its latest models, and artificial intelligence policy advocates warn the move signals the White House is taking an "ad hoc" approach to AI regulation that could hurt innovation. Anthropic disabled access to its newest Fable and Mythos models Friday after receiving a federal export control order requiring it to block foreign nationals from using them. The rare move caught the tech industry by surprise and stoked concerns the directive could set a precedent for how much influence the government can have on AI development and release. "When ad hoc executive actions replace clear standards, America risks surrendering its lead in AI and allowing genuinely dangerous technology to be deployed," said Brad Carson, president and co-founder of Americans for Responsible Innovation, a nonprofit that has been a vocal advocate for stronger AI guardrails. Model takedowns Anthropic revealed late Friday it had pulled its latest models, Fable 5 and Mythos 5, just three days after their release. The models are both based on the Mythos model that the company initially opted not to release to the broader public amid concerns it could supercharge hacking capabilities. Fable 5 was released to the public with safeguards in place meant to protect against uses the company considered dangerous, while a new version of Mythos was provided to a small group of cyber-defenders and infrastructure providers with fewer guardrails. The research community initially slammed Anthropic for intentionally limiting the new models' responses if they suspect users are working on AI research. But the focus shifted to the export controls directive by the end of the week. Several outlets reported Amazon CEO Andy Jassy had reached out to the administration Thursday to flag concerns about a potential method of "jailbreaking," or bypassing, the guardrails on Fable, prompting efforts to get the company to voluntarily pull the models. Amazon is invested in Anthropic. The Hill has reached out to Amazon for comment. Anthropic pushed back, arguing a "narrow potential jailbreak" should not be a cause for pulling a model. It sought to distinguish this from a universal jailbreak, in which an individual can broadly bypass a model's safeguards, and suggested that "perfect jailbreak resistance" is not currently possible. "If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers," the company wrote. "As we have stated publicly, we believe the government should have the ability to block unsafe deployments, as part of a statutory process that is transparent, fair, clear, and grounded in technical facts," it added. "This action does not adhere to those principles." Anthropic staff met Monday with officials at the Commerce Department, an administration official told The Hill. White House's mixed messaging A source close to Anthropic said the Trump administration gave the company 90 minutes to pull Fable, without providing any prior indication about a threat to national security. AI policy figures criticized the unexpected directive, arguing Fable's takedown exemplified a "licensing regime," just weeks after the White House clarified it only supports optional government testing and limited oversight. "AI is licensed now, but the requirements change constantly and are always a secret, even to the administration itself, which will discover the rules spontaneously in real time as it reacts to events," Dean Ball, a co-author of Trump's White House AI Action Plan, wrote in a social media post on Monday. Mandatory government testing is off the table for most AI firms, which warn it could slow down U.S. development and hurt competition with China. The Trump administration has largely agreed, but the release of more advanced models like Mythos forced the White House to consider voluntary model testing. The Mythos release was followed by weeks of debate over how the White House should handle the cybersecurity risks of newer AI models, as the administration tries to balance AI safety concerns with its longtime commitment to light-touch regulation. Trump signed an order earlier this month laying out a voluntary testing process in which AI labs can provide the government with their models up to 30 days ahead of release to test for certain risks. While the administration emphasized testing was not mandatory, some predicted at the time these assurances would not be enough. "Forget 'voluntary,' forget 'permissionless,'" Ball said, adding, "The government wants to apply its force to frontier AI; that much is clear. It wants to make the industry submit." Tensions reignite between the White House and Anthropic Trump allies criticized Anthropic's resistance to taking down its models, suggesting the decision runs counter to the company's safety commitments that have been at the heart of previous disputes with the administration. Former White House AI and cryptocurrency czar David Sacks said in a post on the social platform X on Saturday that the administration has "been very surprised that Anthropic hasn't wanted to cooperate with a reasonable safety request." "Anthropic's reaction is very much at odds with their branding and ethos as a safe AI research community," he added. Anthropic became embroiled in a highly public feud with the Pentagon earlier this year over the safety guardrails in its contract terms. Defense Secretary Pete Hegseth ultimately designated the company a supply chain risk, which the AI firm is challenging in court. Amid this Fable debacle, Hegseth touted this decision, writing in a Saturday social media post: "Every passing day proves why that was the right move." Anthropic accused the Pentagon and Hegseth of retaliating against the company for political differences earlier this year, and some online suggested last weekend's Fable order built upon this. The alleged licensing regime "means also that the rules in practice are stricter and more roughly enforced for organizations the administration does not like," Ball said. One former Trump official rejected these suggestions. Taylor Budowich, who recently departed as Trump's deputy chief of staff, said it is not about a personality and political dispute, but that "Anthropic abandoned its safety guardrails the moment they became commercially inconvenient. It was so egregious their corporate partner -- Amazon -- felt compelled to blow the whistle." The Anthropic debate is "no surprise," said Neil Chilson, head of AI policy at the Abundance Institute, and Adam Thierer, a resident senior fellow with the Technology and Innovation team at the R Street Institute. However, they argued in a Substack post Sunday that "even if you disagree with Anthropic's regulatory strategy, this escalation of government intervention is nothing to celebrate." "It is horrible for the broader AI ecosystem," they wrote. "Continued arbitrary, unexplained deployment of export control authority will make companies slow-walk new models, depriving the public of powerful new tools." "The US government should not hang a Sword of Damocles over every lab's head, with no indication when it might drop or why," the pair added.

How the Anthropic saga could threaten American AI dominance

Why it matters: As the Trump administration shapes its AI regulatory regime in real time, the precedents it sets could reverberate far beyond an individual showdown with Anthropic. Catch up quick: The last month of AI policymaking has been a blur of zigs and zags by the administration. * First Trump delayed an executive order that would have created a voluntary reporting system for advanced AI releases. He said he didn't want to threaten America's lead over China. * A few weeks later, the White House issued a slimmed-down executive order that explicitly barred mandatory government licensing. * Then came Friday: The administration placed Anthropic's Mythos 5 and Fable 5 under export controls, a move critics said amounts to a licensing system by another name. Driving the news: The move against Anthropic -- which came as the Pentagon is already tangling with the company -- has some foreign governments doubting they can depend on U.S. AI. * "The situation we're in collectively right now with Mythos and Fable is something that can happen with overreliance on certain models," Canadian Prime Minister Mark Carney said Sunday. "You'll hear me say this over and over again. It is never a good idea to have one option," Carney said. * Early this month, the European Union launched a "tech sovereignty" initiative to reduce dependence on foreign technology providers, including American AI and cloud companies. It wants to dramatically expand data center and semiconductor production. * ""Europe wants to be in the position to make its own choices, avoiding risky dependencies on single dominant suppliers, one company or one third country," European Commission Executive Vice-President Henna Virkkunen, who oversees tech sovereignty, told reporters. Friction point: The risk is that foreign governments and companies will look elsewhere if they come to view American AI as unreliable. Google DeepMind CEO Demis Hassabis said in January that China is roughly six months behind the U.S. when it comes to frontier, or leading, AI. * "For some applications, Chinese models will be an attractive back-up, especially when they are open-sourced and so can be used with relatively little risk," Anton Leicht, an AI expert at the Carnegie Endowment for International Peace, tells Axios. Yes, but: Overall, no other country -- including China -- is close enough to the U.S. in data center buildout or chip production to give the Trump administration second thoughts, Leicht added. What they're saying: Gartner noted in a Monday report that this was the first time that a government has intervened to block access to an AI model customers were already using -- and warned it probably won't be the last. * "Operational risk can stem not just from a vendor's performance but also from unpredictable government interventions," the technology research firm said. * "The Trump administration is collaborating with AI industry leaders to balance cutting-edge innovation with national security concerns that affect both the United States and our allies," White House spokesman Kush Desai said. "The United States is by far the world leader in the global AI race, and President Trump is committed to ensuring America's technological dominance." The bottom line: The Trump administration is sending mixed signals about its approach to AI regulation -- and it's giving allies abroad the jitters.

Trump's Anthropic crackdown sets off AI alarms for U.S. allies

With the recent crackdown on Anthropic, the White House has given global leaders another reason to panic about their place in the technology race. On Friday, the U.S. ordered Anthropic to deny foreign nationals access to the company's newest artificial intelligence models. The export ban asserted a broad, unprecedented authority over the technology. Until then, conversations in Europe about losing access to U.S. tech -- sometimes posed as a presidential "kill switch" -- were theoretical. To many on the continent, Friday's move underscored the dire need to find alternatives to American AI, and fast. France's Prime Minister Sebastien Lecornu announced Tuesday that the civil service would roll out a tool based on local startup Mistral AI. He added that French company Chapsvision had been selected by the French domestic intelligence agency DGSI to replace U.S. software firm Palantir. "France must have its own tools," Lecornu said. "We cannot rely on the goodwill of certain partners who, as we have seen in recent days, are capable of cutting off access to the Anthropic model." U.S. allies -- already reeling from the Iran war and U.S. President Donald Trump's wavering commitment to NATO -- must now confront a future in which the White House can pull the plug on AI sales abroad as it pleases. The U.S. government has previously used similar export control measures to control access to AI chips. The leaders of the Group of Seven countries included AI as one of the key points of discussion at the meeting being hosted by French President Emmanuel Macron in Evian, France, which runs through Wednesday. A draft statement said the group would "further discuss emerging opportunities and potential risks arising from AI, notably in the financial sector." The latest ban, which hits Anthropic's Mythos and Fable models, affects software that banks, law firms and government offices are rushing to adopt. That decision "materializes a risk that had been on everyone's mind" outside of the U.S., said Mathilde Velliet, who focuses on the U.S. tech strategy as a research fellow at the French Institute for Foreign Relations. Europe, the Middle East and Canada have struggled to create their own AI companies that are of a similar scale to the U.S.-based frontier models, meaning local companies and governments often choose dominant American products that make them more vulnerable to the Trump administration's whims. The U.S. export ban "shows how the U.S. government views Europe: as an enemy, not as a friend and ally," Alexandra Geese, a European parliamentarian from Germany, said in a statement. French presidential candidate for the centrist Renaissance party, Gabriel Attal, declared over the weekend that the "AI war has already begun." His rival on the far...

IndAI sovereignty, a fable: Frontier AI is now a geopolitical tool -- India must leverage its market accessibility

India faces a critical juncture as powerful AI models are restricted by the US. This move impacts the security of India's digital infrastructure. The nation must prioritize developing its own AI capabilities and ensuring sovereign control over its technology. At the time that Anthropic unexpectedly implemented the US government's export control directive on its powerful AI model Fable, I was using it, trying to strengthen security for my vibe-coded apps one by one. With one sweep, Fable's usage was stopped for anyone who was not a US national. Fable is a restricted version of Anthropic's powerful Mythos AI model, which has found security vulnerabilities in well-maintained software containing code that has been security-tested for as long as 16 yrs. These are called zero-day vulnerabilities: security loopholes and gaps that remain undiscovered, and can be used to launch attacks on software before they're identified and fixed. One well-known zero-day vulnerability was in WhatsApp when Pegasus used it to infiltrate devices couple of years ago, which was subsequently patched. Fable was restricted, and meant to be used to identify and fix vulnerabilities, not exploit them. By banning its use for foreign nationals, the US has prevented the rest of the world from securing its critical software and defending itself. So, the model that was restricted is not uniquely dangerous. This affirms sovereign control over frontier AI models, and that developing frontier models is now a national security consideration. There is precedent for this. During Trump 1.0, the US briefly restricted security updates for Android OS on Huawei devices. China's response was to develop its own mobile OS, HarmonyOS. Most Indian handsets run Android, and US tech companies power significant infrastructure in India. The rest of India needs to follow Kerala's approach and adopt open-source technologies, especially when it comes to critical deployments. So, how sovereign will the infrastructure that OpenAI-TCS builds be, especially given OpenAI's deal with the US Department of War? We must be clear that sovereign AI means that deployments are on Indian soil, and all control resides with Indian companies, with no external kill switch. When Anthropic announced a partnership with TCS, the same day it pulled Fable, Dario Amodei's 'commitment to India, our second-largest market, rang hollow. When Anthropic launched Project Glasswing, its effort to 'secure the world's most critical software', partners listed were Amazon Web Services (AWS), Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorganChase, Linux Foundation, Microsoft, NVIDIA and Palo Alto Networks - US companies, US interests, US infrastructure. Anthropic had the opportunity to give access to its partners Infosys, which runs Finacle in banking, and TCS, which runs a host of India's digital infra, including passport and tax systems, with access to highly sensitive personal data that's now at risk. We need to recognise that for companies like ChatGPT and Anthropic, India is a market for AI products, and nothing more. The Nishikant Dubey-chaired parliamentary standing committee on IT should seek answers from Anthropic's representatives in India on what the company's obligations are to Indian infrastructure operators - whether Project Glasswing access will be extended to companies running India's critical systems, and what conditions govern any future access suspension. India's cybersecurity agencies are out of their depth. CERT-In's response to the panic that spread after Claude announced Mythos AI was to hold meetings with Indian infrastructure providers to ask them what they could do to strengthen cybersecurity against Mythos. It is next to impossible to detect and address zero-day vulnerabilities, and bureaucratic checkbox exercises serve no purpose. We need less theatre and more competence. Indian committees often look for big names and credentials, not competence. These are specialised areas where expertise matters. Nandan Nilekani's suggested focus on India building AI applications over developing its own LLMs should alone disqualify him from any key role. Let's also not forget that Aadhaar was so badly built that read-write access to the database was being sold on WhatsApp groups for ₹500. We need to build AI capacity. Atal Tinkering Labs (ATL) is a laudable initiative for encouraging school students to build tech. India AI Impact Summit in New Delhi in February did well in promoting AI diffusion. We need more such summits to encourage students. Importantly, we need interventions at university and industry levels. Allow universities to monetise research, and professors and students to start companies with university support. Build a marketplace that rewards innovation and expertise. Consider launching a special visa, with perks, to attract AI talent to the country. India also needs to strengthen IndiaAI Mission, provide hardware and encourage competition in AI model development, and not just put all its eggs in the Sarvam basket. India is the second-largest market and largest available market for US tech companies in terms of user base. They have significant investments and customers here. We need to use market access as a mechanism to build leverage with the US, especially since it's clear that the win-win era of geopolitics is over, and that access to frontier technology is being used to exacerbate vulnerability. If that means that Anthropic won't get access to the Indian market if it doesn't give access to the latest frontier models, so be it. For those looking to strengthen their deployments, OpenRouter has made available a version of the Fable AI model. Critical infrastructure providers should use that to secure their deployments in the absence of access to Fable from Anthropic.

G7 Leaders Voice Concerns Over US Cutting Off Access to American AI Models

"For India, AI means 'All Inclusive" and the fact remains that any technology can lead to progress only if democratised, the Prime Minister has said India joined G7 leaders like French President Emmanuel Macron in raising concerns over the United States unilaterally cutting off access to American AI models without warning. Obviously, what set it off was the Trump administration's blockage of Anthropic from exporting its newest Mythos 5 and Fable 5 models over national security. Prime Minister Narendra Modi, who met President Trump on the sidelines of the G7, told VivaTech 2026 - Europe's largest technology event - that "For India, AI means 'All Inclusive" and highlighted that technology can lead to progress only if democratised. Modi said India has always viewed cyberspace as a global public good, Prime Minister emphasized that democratic countries must have access to AI models that can secure their critical information infrastructure and help them deal with cyber threats. However, Macron was more direct at the G7 conclave where he warned its leaders and top AI executives including Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman as well as Trump that if the US "from one day to the next can turn off the switch", it could harm European economies and customers but also damage AI firms themselves. In fact, leaders were unanimous in their views that the Anthropic episode last week that the Trump regime had used dubious claims to make Anthropic withdraw its latest models from the global market. Readers would recall that the White House had reacted to Amazon's flagging that some safety guardrails could be bypassed. This was furiously countered by cybersecurity experts who claimed that the same issues cited by the government were present in multiple models that remain freely available. This included models from Anthropic's main rival OpenAI. This episode exposed a real risk that several global governments and enterprises operating in their countries felt. Indian companies were among the first to respond with Zoho CEO Sridhar Vembu saying the US moves shows the growing link between technology, national security, and economic sovereignty. He took to X and declared that "globalisation is dead and Bharat must find her own way ahead." Sarvam co-founder and CEO Pratyush Kumar echoed these sentiments stating that countries cannot afford to mistake access to cutting edge AI systems for true technological ownership. "Fable 5 ban is a good instigation for more people to engage in recognising the need for sovereignty," Kumar said via his X handle. The Trump administration's devil-may-care outlook hasn't gone down well with governments who believe that any company or government that builds on US AI infrastructure has to reckon with the whims and fancies of one single leader in an administration who can revoke access overnight and never actually mention the reasons for it. In fact, Modi also referred to Trump's blocking of Anthropic's latest models stating that democratic countries must have unfettered access to top AI models for protecting their own critical infrastructure. In fact, earlier in the day, Anthropic CEO Dario Amodei had echoed similar sentiments of creating a global partnership for AI research and expansion. The G7 leaders also discussed the creation of a "trusted partners" regime that would provide access for non-US nations to advanced AI models. However, what it still does not bring to the table is how the world can tackle issues such as the one Trump created recently by switching off supply at the blink of an eyelid. Not to mention that the President is also trying to ensure that all future models are first tested by the White House before release to the public, indicating that a global AI hegemony has already developed in more ways than one. Maybe, it is time to seek out DeepSeek! However, Macron believes that the trusted partnership program itself could do the trick. The G7 believes that such a move could actually open a path around US restrictions as these partners could be countries or companies that allow usage of their models to develop stronger cybersecurity defences. Given that the US and China together control roughly 90% of the global AI computing infrastructure (per data released by AI research firm Epoch AI), how far this partnership idea can go is anybody's guess. What is making the G7 leaders squirm is the erratic and threatening behaviour of President Donald Trump in his second term at the White House. Thus far, he has unilaterally raised tariffs before bringing them down himself, asked Denmark to surrender Greenland to the US, ring-fence US technology companies on grounds of national security and wage a war against Iran for purposes which the world still hasn't fully understood. Just as Indian business honchos are calling for sovereign AI, politicians across Europe are also asking companies there to develop their own AI models, with several reactions focusing on Mistral, the most credible AI contender in the continent. Maybe Prime Minister Narendra Modi should review this statement by Aidan Gomez, CEO of Canadian AI company Cohere. "Digital sovereignty is not just about market competition or any one company or nation. It's about who controls the foundational technology that will shape our economic security and national sovereignty for decades to come." Trump might say that India has nothing to worry till he is around, but the reality is that he has his finger on that Kill Switch that may starve others of American AI prowess. It is indeed time to act and the likes of Sarvam and Zoho must step now, with government support.

Claude Fable 5, Mythos banned: Indian techie shares Indian jugaads one can use to bypass restrictions to access AI tools

New US restrictions on advanced AI models from Anthropic have caused a stir in India. Indian tech enthusiasts are sharing 'jugaad' solutions to bypass these limitations. This situation highlights India's reliance on foreign AI platforms and fuels calls for stronger domestic AI development. The focus is now on building India's own AI infrastructure to meet local needs and compete globally. Anthropic's latest artificial intelligence models Claude Fable 5 and Mythos 5 have triggered a fresh debate in India after access restrictions linked to US government export controls left several international users unable to use the advanced AI systems. As discussions around the ban spread online, Indian developers and technology enthusiasts started sharing their own "jugaad" ideas, while many argued that the situation highlights the need for India to build stronger homegrown AI capabilities. Also Read: 'I lost my job to AI': 24-year-old Meta data scientist says layoff was more relief than pain, big tech jobs are no longer safe The restrictions have created a new conversation around access to advanced technology, global AI competition, and whether countries should depend heavily on foreign AI platforms. Why Claude Fable 5 and Mythos 5 access is banned Anthropic, the US-based company behind Claude AI models, recently introduced Fable 5 and Mythos 5 with improved capabilities in areas including general AI tasks and cybersecurity-related applications. However, concerns around possible misuse and national security risks led to tighter controls from the US government. Under the new restrictions, access to these AI models has been limited mainly to US citizens. The move reportedly even affected well-known AI researcher Andrej Karpathy, who recently joined Anthropic, as he was unable to access the models because he is not a US citizen. The development has raised concerns among global developers who rely on advanced AI tools for research, coding, startups and innovation. Indian techies reacts with 'jugaad' solutionsSoon after the restrictions came into discussion, social media platforms in India saw a wave of reactions from developers and technology users. Many Indian users shared possible workarounds involving VPNs, proxies, overseas connections and other methods. The conversation quickly turned into a familiar debate around the Indian culture of finding solutions when faced with restrictions. One viral reaction summed up the mood among many users: "Do they really think they could stop Indians?" Also Read: Sundar Pichai tells Stanford graduates 'your choices will shape your journey' after students walk out during speech Several users pointed out that software restrictions have historically not prevented determined communities from finding alternative ways to access technology. At the same time, experts warned that stronger identity checks and compliance measures could make such access more difficult in the future. AI restrictions raise questions about technology dependenceThe Claude restrictions have also brought attention to a larger issue, dependence on foreign AI systems. India has one of the world's largest pools of software engineers, developers and technology entrepreneurs. However, many startups and professionals still rely on AI models created by companies outside the country. The latest development has renewed calls for India to accelerate investments in domestic AI models, computing infrastructure and data centres. Is it time for India's own AI push?India has already started focusing on building its AI ecosystem through initiatives such as the National AI Mission and increased collaboration with global technology companies. Technology leaders believe that the country needs stronger AI infrastructure that can address local requirements, including Indian languages, agriculture, healthcare, education and governance. The argument is that instead of depending entirely on foreign AI platforms, India should create models designed around its own data, challenges and user needs. From jugaad to innovation: India's AI challengeThe debate around Claude Fable 5 and Mythos 5 access is not only about bypassing restrictions. It has also become a discussion about India's future role in the global AI race. While some developers are looking for short-term solutions, others believe the bigger answer lies in creating powerful Indian AI systems. As countries compete for leadership in artificial intelligence, the latest restrictions have added urgency to India's push for technology independence. The message from many in the tech community is clear -- access matters, but building the next generation of AI may matter even more.

US Blocks Fable 5: AI Safety or Geopolitical Kill Switch?

The US government has ordered Anthropic to suspend access to its flagship AI models, Fable 5 and Mythos 5, under an export control directive issued on June 12. Anthropic said the order applies to "any foreign national, whether inside or outside the United States", including its own foreign-national employees. The company has consequently disabled both models for all customers worldwide, while access to its other AI models remains unaffected. Anthropic disputes the rationale behind suspension: Anthropic said the government linked the directive to concerns about a potential method of bypassing, or "jailbreaking", Fable 5's safeguards. However, the company said it was not given "specific details" of the national security concern and had received only "verbal evidence of a potential narrow, non-universal jailbreak". According to Anthropic, the demonstrated technique uncovered only "a small number of previously known, minor vulnerabilities" that were "relatively simple", adding that "other publicly-available models are able to discover them as well without requiring a bypass". The company defended Fable 5's safety measures, stating that its safeguards are "substantially more effective than those of any previously deployed model" and that no tester had found "a universal jailbreak". It also reiterated that "perfect jailbreak resistance is not currently possible for any model provider" and said it had adopted a "defense in depth strategy" combining safeguards, monitoring and 30-day data retention to detect and mitigate attacks. Company says other frontier models have the same capabilities: Anthropic argued that the reported capability was "widely available from other models (including OpenAI's GPT-5.5)". While complying with the order, it said it disagreed that "the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people", warning that such a standard could "essentially halt all new model deployments for all frontier model providers". How people are reacting online: * Sovereign AI wake-up call: Sarvam AI founder Pratyush Kumar said users should not "confuse access with ownership, or adoption itself as advantage", warning that organisations leveraging technologies with "external control loops" must accept that they are vulnerable. He argued that the incident strengthens the case for "Sovereign AI", where India can "use and improve AI systems within their own perimeters". * Former MP Vijayasai Reddy similarly called the move a "wake-up call", warning that India "can have its access restricted at any time" and urging the government to accelerate investments in domestic frontier AI and compute infrastructure. * AI becoming a strategic asset: Public policy commentator Ajay Purushotham Shetty argued that the export controls underscore an emerging reality that "frontier AI is now a strategic asset". He said "Software Swarajya is no longer optional" and warned that countries that do not control their own AI infrastructure risk losing strategic autonomy. According to Shetty, "those who control intelligence infrastructure will shape the future", making domestic research, compute and model development a national priority rather than a commercial choice. * How AI regulation could look like at scale: Aaron Levie, CEO of Box, framed the episode as an "early peek into what AI regulation would end up looking like at scale". He argued that the government would effectively gain "sole discretion over when a model can be released to the public". Levie warned that such a framework could result in a "backlog of AI releases" and cause progress to "dramatically slow down". While stressing that "AI safety is incredibly important", he argued that restricting models at the foundation layer rather than targeting harmful applications risks stifling innovation across the sector. * Hurts everyone by making all systems vulnerable: Nikhil Pahwa, founder and editor of MediaNama, questioned the logic of restricting access to a model that many developers were using to identify and patch vulnerabilities. He noted that Fable "comes with security safeguards to prevent misuse" and said he had personally used it to "identify and plug security issues" in applications. Pahwa argued that "the US government refusing to let non-US nationals protect themselves is deeply problematic in an interconnected internet world", particularly when vulnerabilities in foreign systems can ultimately affect American users as well. "This hurts everyone by making all systems vulnerable," he wrote. * Open-source and market alternatives: Entrepreneur Upamanyu Acharya argued that the suspension is unlikely to slow AI adoption because users will simply switch providers. He described the development as "bullish for open weights, not bearish", noting that when "one frontier lab" is removed, demand "reroutes to the next lab". Drawing on his own usage patterns, Acharya said he runs significantly more workloads on open-weight models such as DeepSeek because they deliver roughly "90% of the output at a 1000x less the cost". * The real sovereignty problem: While many commentators called for sovereign AI models, Acharya pointed out that India faces a much deeper challenge. "That, not the Fable ban, is the real sovereignty problem," he wrote, referring to the fact that AI spending largely flows overseas. He argued that model weights are relatively easy to replicate, claiming that "a sovereign model is the one layer of this entire stack we could have cloned in a weekend". The real dependencies, he said, lie in foreign-controlled infrastructure: "power, copper, fabrication, ports, rare earths", where India has "zero position". "Every input I pay for sits outside the country," he added. * Geopolitics of frontier technology: Entrepreneur Sandeep Manudhane described the suspension as evidence that frontier technologies are increasingly being "geo-weaponised". He characterised the move as the US government exercising a technological "Kill Switch" and warned that countries without domestic capabilities remain exposed to foreign political decisions. Manudhane argued that India is already a "near-total American digital colony" and suggested that similar restrictions could theoretically affect other critical technologies in the future.

Kishore Lulla says Anthropic restrictions expose India's AI dependence; calls sovereign AI a 'national necessity'

Eros Innovation Founder and Executive Chairman Kishore Lulla has said that the US government's decision to restrict foreign access to Anthropic's most advanced AI models should serve as a wake-up call for countries that remain dependent on overseas AI platforms and infrastructure. The comments come after Anthropic suspended access to its Fable 5 and Mythos 5 models following a US government directive that classified the systems under stricter export controls and barred access for foreign nationals on national security grounds. Anthropic has said it complied with the order while disagreeing with the government's assessment. Reacting to the development, Lulla said the move demonstrates the risks of relying on foreign-controlled AI systems for critical technological and cultural capabilities. "I have said this before and I will say it again - this was inevitable. When a nation's cultural expression depends on foreign platforms, foreign models and foreign infrastructure, those platforms will always make decisions that reflect their own priorities, not ours. Claude restricting access to Indian users is not a surprise. It is a consequence of a structural dependency that India must urgently address. This is precisely why we built the Eros LCM family as a sovereign AI initiative, recognised by MeitY, anchored at IIT Madras, trained on Indian data that we own and control. India cannot afford to be a consumer of AI built elsewhere. Every time a foreign platform restricts access, changes terms, or simply decides that India is not a priority, we are reminded of what is at stake. Sovereign AI is not a policy ambition. It is a national necessity. India has the data, the talent, the institutions and the government commitment to build it. The question is urgency. I hope this moment accelerates it." According to reports, the US order requires Anthropic to halt access to Fable 5 and Mythos 5 for foreign nationals as Washington expands export-control measures around frontier AI systems, citing national security concerns. The move has triggered debate across the global technology ecosystem about access to advanced AI models and the strategic importance of domestic AI capabilities. Lulla said the development reinforces the argument that countries such as India must invest in indigenous AI models, datasets and infrastructure that remain under domestic control, rather than relying exclusively on technologies whose availability can be altered by foreign governments or corporations.

Thanks, Trump-ji for the push we needed

US attempts to control advanced AI model exports face challenges. This may push nations like India to develop their own AI capabilities. India's strong ties to the US software sector mean it must consider diversifying its AI sourcing. Building sovereign AI offers strategic and economic advantages. India can achieve domestic AI resilience with its talent pool. The Trump regime's attempt at controlling the export of Anthropic's most advanced AI models, Fable 5 and Mythos 5, may not play to script. The San Francisco-headquartered AI company threatened to disable the models if it is asked to keep access confined to US nationals. The company has a history of challenging the government. It has refused to allow the US military to use its AI models for autonomous warfare. It has also delayed the public release of Mythos, which can detect security flaws, until companies building critical software run conclusive tests with the AI model. This infrastructure is international, and denying access to non-Americans runs counter to Anthropic's original intent. However the developments turn, expansion of US export controls from hardware to software will force countries to seriously rethink AI sourcing. India is particularly affected because of its strong linkages to the US software industry through manpower and data. Indian engineers help build AI models, which then train on Indian data. Denial of access to frontier technologies will provoke diversification as well as localisation. India anyway needs sovereign AI because of its cultural diversity and need to protect data. It also has a talent pool that can drive indigenisation. But it will have to depend on advanced US semiconductors to build its foundation AI models. Domestic AI resilience is within the reach of many countries, including India. Locally-owned AI infrastructure, data and models can scale up surprisingly fast. Strategic and economic benefits of sovereign AI are already shaping market behaviour, with a little help from the government. This has a bearing on policy responses of AI leaders as the race becomes crowded. The hyperscale US model requires a global market for its AI innovations. The state-led Chinese model also depends on overseas customers. India needs to keep its sourcing options open as it builds domestic AI capability. It pays to have friends all over in this game.

Globalization is Dead: India Must Shrug Off LLMs and opt for Open-Source Small AI Models

Many tech experts are seeing the move as a wake up call for Indian Tech industry to wake up and build open-source small language models In the new world of technology hegemony wrapped up in national security concerns like we saw with the Trump administration getting Anthropic to remove its latest AI models from the shelf, India needs to tighten its AI belt, remove development cobwebs and get down to prepare its own open-source AI models. For long a section of tech experts have been voicing the need for data sovereignty, asking whether it would be appropriate as one of the world's largest AI markets to rely on technology built and controlled elsewhere. That the move came without hours of Anthropic partnering with India's IT services giant TCS for enterprise AI deployments further accentuates the need for a debate now. Under this deal, the Indian company was to create a business unit focused on deploying Anthropic's AI models to its customers while also gaining early access to new model releases. Among the first to call this out was Zoho CEO Sridhar Vembu who said Anthropic being forced to withdraw their new models shows the growing link between technology, national security, and economic sovereignty. He took to X and declared that "globalisation is dead and Bharat must find her own way ahead." In a separate post, Vembu described Dr. Abdul Kalam (formerly President) as a visionary "way ahead of his time" who saw the need for sovereign tech back in 1998. "We have to work hard to catch up now, but it is doable," he said while noting that the private sector must make R&D a big priority with government only providing additional assistance. Vembu's response came to noted economist S Gurumurthy's post where he said America is doing what Dr. Kalam wanted India to do in 1998. "After Pokhran II blast US-West sanctioned India. Kalam told me (which I have recorded many times) to persuade PM Vajpayee (who I was close to) to ensure the sanctions continue for our tech to develop," Gurumurthy said. As we noted in an earlier post, Anthropic's predicament is far more serious than merely linking it to their future IPO impact. Media reports in the US said US Commerce Secretary Howard told Dario Amodei of Anthropic that companies would require a special licence to distribute the Mythos and Fable 5 models to "all destinations worldwide" and that failure to comply would result in criminal and civil penalties. Which is why tech experts who had earlier forewarned us from falling head-over-heels with the US companies are now sitting up and saying "I told you so" and hoping that this development triggers a much-required debate among tech founders, investors and policy experts. Some experts this publication spoke with said it was a "necessary wake-up call." That India is the second largest market for both Anthropic and OpenAI is no secret. Now add Google into this mix and it is quite clear that none of these companies can ignore India. But, a trigger-happy President and his penchant for tariffs as a means to conquer is another proposition. Which is why the statements of Vembu and others make a lot of sense. Moreover, these concerns have been building for some time. The recent decision of Opendoor to shut its India global capability centre citing AI as the cause is one such. CEO Kay Nejatian said the push to bring operational work closer to the customers in the US and a shift towards smaller AI-native teams pushed them to shut down and give pink slips to 250 people in India. Which is also why Vembu's comments make sense where he notes that in the modern age, technology is the "ultimate weapon" and it would do India good to embrace smaller and open-source models. "What can our government do right now? Ensure that orgs in India embrace smaller models, both Indian and Chinese open source ones," Vembu said. Of course, others like Mohandas Pai wanted to go broader and centralised. He called for a national AI strategy (as if we do not have one already!) where government must increase investments on AI, compute infrastructure, and deep technology. He wanted a National AI Taskforce headed by Narendra Modi and his former boss Nandan Nilekani as chairman. Meanwhile, Sarvam co-founder and CEO Pratyush Kumar said countries cannot afford to mistake access to cutting edge AI systems for true technological ownership. "Fable 5 ban is a good instigation for more people to engage in recognising the need for sovereignty," Kumar said via his X handle. Writing for The Print, KBS Sidhu is a former IAS officer who retired as Special Chief Secretary, Punjab, noted that "India's response to Claude Mythos must begin not with a committee, but with a phone call. The government should formally approach Anthropic for inclusion in its trusted-partner programme -- the same access extended to Microsoft, Apple, and Google." The rationale is self-evident: no country in the world has built a digital public infrastructure of comparable scale or systemic importance. UPI, Aadhaar, NPCI, the government email stack -- these are not corporate assets. They are sovereign infrastructure serving citizens. If Mythos can find vulnerabilities in Windows and Chrome that went undetected for decades, it can find vulnerabilities in India's payments and identity architecture, he said. India must also use its international weight. As a G20 member with proven convening capacity, India should champion a multilateral framework on offensive AI governance. Cyber threats are jurisdictionally blind. An AI-generated attack on NPCI's switching layer cares nothing for the nationality of the server it originates from. A governance architecture that mistakes patriotic branding for security assurance is no architecture at all, the former IAS officer said. It looks like Trump has given yet another wakeup call to India, just as he did with his tariff diplomacy. The question is whether India will take the cue and go all out for tech sovereignty or would it continue to kowtow to the United States and its Big Tech hegemony?

Tech denial's come home: The Anthropic shock is India's wake-up call

Subimal Bhattacharjee Commentator on digital policy issues Last Friday, Anthropic switched off its two most capable models, Fable 5 and Mythos 5, for most customers. It did so not by choice but after being arm-twisted by a US government export-control directive, citing national security issues, that bars access by any foreign national, whether inside or outside the US, including Anthropic's own foreign-national employees. Because Anthropic couldn't cleanly separate US citizens from non-citizens, compliance meant disabling the models altogether. So, a frontier AI tool that was launched days earlier went dark within hours of a letter from the US commerce department. This also disrupts the Anthropic-led expanded Project Glasswing - uniting major tech firms like Amazon, Google, Microsoft and Apple - to combat AI-driven cyber threats that India would have benefited from. If one strips away specifics, the episode is a signal flare. For a decade, technology denial has crept from the periphery of sanctions on adversaries (like Huawei) to the centre. 2025 AI Diffusion Rule sorted the world into tiers and, for the first time, treated model weights as controllable munitions. Friday's directive is the next turn of that screw. The frontier of control is no longer geographic but the individual. What is alarming is the precedent it sets. If a security concern can vapourise a publicly launched model overnight, every other deployment now carries that contingent risk. That should unsettle every government betting its AI future on the US stack, India included. India has wagered heavily on tech diplomacy with the US - TRUST (Transforming the Relationship Utilising Strategic Technology) initiative launched by Trump and Modi in February 2025, the Pax Silica supply-chain compact, and India-US AI Opportunity Partnership signed this February. The framing is seductive with India as 'trusted partner' anchoring AI infrastructure outside East Asia's chokepoints. The latest Fable 5 episode, however, exposes the fine print - conditional trust that's revocable overnight. If Washington can turn off the lights on a model for its own engineers, an Indian developer's access is a privilege, not a right. China will read the same event and harden its conviction that sovereignty must be built at home, because any spigot the US controls is political. Even after the US cleared Nvidia's second-tier chips for Chinese buyers last December, Beijing has kept pouring capital into an indigenous stack. Middle powers are improvising, too. France, hosting this week's G7, is courting a 45 bn SoftBank build-out and branding itself as Europe's AI hub. The EU is chasing 'sovereign' compute. Japan and South Korea are nursing domestic foundries, while the Gulf trades investment for chip allocations. None yet rivals the US frontier. But the Fable 5 shock is exactly the kind of thing that accelerates the hunt for alternatives, the very fragmentation US strategists fear. The timing is almost theatrical. The 3-day G7 summit convenes in Evian-les-Bains from today, where Macron hopes to nudge AI cooperation forward, even as governance language is diluted and Washington presses its 'American AI technology stack' against Chinese rivals. India is attending as a guest, with lot of tech diplomacy work happening at the back. 120 Indian deep-tech startups are there in Nice since Sunday for the 3-day Bharat Innovates 2026, spanning semiconductors, quantum and space. Narendra Modi will be a speaker at the annual VivaTech conference in Paris later this week, where India is 'AI Country Partner' and is expected to unveil its MANAV (Moral, Accountable, National Sovereignty, Accessible, and Valid) governance framework. Founders in Nice are courting global investors at the exact moment the US has shown conditional tech control without warning. So, is 'trusted partner' now a probationary perch? Does anchoring to the US stack lock India into a dependency a future directive could throttle in an evening? The answer is not to walk away. Compute, capital and frontier models still flow predominantly from the US. What India should do is refuse exclusivity. India's strongest play is as that it's embedded deep enough into the US ecosystem to extract real infrastructure. It's sovereign enough in talent and standards to bargain, and open enough to France, Japan and its own stack to avoid single-point capture. The ecosystem is ramping up but GoI has to foster this further. Because technology denial has come home.

Big yard, iron fences: Why India must chase AI sovereignty now

Jay Vinayak Ojha The writer is project fellow, Vidhi Centre for Legal Policy When the US imposed strict controls on the export of advanced GPUs, the computing backbone of AI, to China in 2022, then-NSA Jake Sullivan declared that the intention was to build a 'small yard with a high fence'. The state would guard a few critical technologies jealously. But beyond that narrow enclosure, the free commerce of the world would continue undisturbed. US commerce secretary Howard Lutnick's order last week directing Anthropic to deny its latest Fable 5 model (and Mythos 5) to all non-citizens -- including engineers who built it -- marks a dramatic enlargement of that yard. Should the directive hold in all its severity, the fences will become iron walls. The decision lays bare three developments that will survive any potential reversal: uState, the supreme actorVarious competing ideologies, including those evangelising 'long-termism' and 'effective altruism', seem to believe that the mild-mannered, soft-spoken, bespectacled lot in Silicon Valley -- not the gruff general or clueless politician -- should be entrusted with the fate of the world. Over the course of this year, this notion has been mugged by reality. When in February, Anthropic declined Pentagon's demands to strip safety safeguards from its Claude AI model to be used for any 'lawful use' by the government, it cited red lines on domestic surveillance and autonomous weapons. Accordingly, the company was struck from the list of approved government contractors, and branded a supply-chain risk. While this regime can rarely be accused of acting with proper purpose, it can hardly be blamed for being peeved at the company. Last week, Anthropic CEO Dario Amodei warned that AI will become the 'dominant source of military and economic power'. But a firm can't declare its tech to be the decisive instrument of national survival, deny it to its own nation's defence services, and then be surprised at the state's hostility towards it. This tension between universalist aspirations of technologists and realities of a hard-power world dominated by states will have to be confronted honestly by all countries, including India. In every previous generation of technological change, big tech has adjusted to needs of the US state. There is no reason to believe this time will be different. v Resurrection of US hard power The organising myth of the post-Cold War era was that US ideas and ideals would simply spread openly and peacefully around the world through the internet. What this ignores is how the Cold War itself was waged, and won. The US successfully banded together in a coordinating committee (CoCom) with its allies in Western Europe, Britain and Japan to successfully deny the Soviet Union and Eastern bloc countries access to the technologies that would dominate the computing revolution: microelectronics and telecommunications. This preserved the West's qualitative 'lead time' in the arms race. The US and its allies forewent immediate profits in pursuit of a grander strategic pursuit. It appears that the Trump regime today is willing to do the same. India will have to adapt to a world where money will no longer be enough to buy critical technologies. w Weaponised interdependence The idea of weaponised interdependencedeveloped by political scientists Henry Farrell and Abraham Newman posits that the power seated at the central node of a network -- finance, information, and now computation -- can chokeoff all who depend upon it. The latest Fable 5 order is that choke point made manifest. To raise an economy and armed forces upon AI models that temperamental foreign officials may switch off at will is extremely imprudent. India has shown before that it can punch above its weight in the contest for strategic technology through coordinated national effort. Cut off from all assistance in the wake of the first Pokhran test in 1974, India's atomic scientists managed to build an entirely domestic nuclear ecosystem. Isro and UIDAI also stand as testaments to what can be achieved through coordinated effort. This latest development in the US simply drives home the fact that it's past time to kick IndiaAI Mission into high gear, with an empowered governance structure and frontier-AI sovereignty as the explicitly stated goal.

Anthropic ban: Sarvam AI's Pratyush Kumar warns against reliance on foreign models

A US ban on advanced AI models has highlighted the vulnerability of foreign reliance, according to Sarvam AI's Pratyush Kumar. He urged Indian firms to embrace sovereign AI, building systems domestically. For AI researchers and engineers, he said the ban sets a precedent in which talent will increasingly be seen as aligned to national interests rather than the companies. After last week's US government ban on Anthropic's frontier models citing security risks, Sarvam AI cofounder Pratyush Kumar said Indian AI firms and users have been operating under a dangerous illusion. The Trump administration on Friday ordered Anthropic to cut off access to its two most advanced models, Fable 5 and Mythos 5, for all foreign nationals, whether inside or outside the United States. Anthropic complied by taking both models offline globally. This is the first time the US has issued an export control on an AI model, and the access was revoked with no warning. "For AI users, it is clear that you should not confuse access with ownership, or adoption itself as advantage. And if the most significant tech differentiator you are leveraging has external control loops, then you have to accept you are vulnerable," he wrote in a post on X. For AI researchers and engineers, he said the ban sets a precedent in which talent will increasingly be seen as aligned to national interests rather than the companies. For AI labs, Kumar argued the episode points toward a deliberate two-tier market: general-purpose models distributed freely to build adoption, with frontier models kept behind a gate. "Democratised AI sucks in all the data liquidity of the world which is locked in higher margin frontier offerings," he wrote. Sovereign AI Kumar said this is a good opportunity for Indian companies to turn to sovereign AI, with nations and companies building, running, and improving AI systems within their own borders, rather than depending on foreign labs that answer to foreign governments. Kumar detailed that Sarvam has trained models at scale on roughly 3,400 Nvidia H100 GPUs and has brought India's first Blackwell cluster online, targeting tens of megawatts of compute on Indian soil by 2027. He added that its flagship model, Sarvam 105B, is described as India's first sovereign foundation model built from scratch. Larger models focussed on coding, agents, and security are in development. "Bottomline, irrespective of hit jobs by foreign media or naysayers of various shades, India has the talent, the scale of the economy, the ambition amongst people for a better life, and a government that understands the imperative for technology innovation," he concluded in his post.

America's shutdown of Anthropic's AI models is not a surprise. And India has seen it coming for 30 years

The US government's export control order on Anthropic's advanced AI models, blocking non-American nationals, highlights a long-standing Indian pattern of adopting foreign technology without developing indigenous alternatives. This "kill switch" event underscores the vulnerability of nations dependent on foreign tech infrastructure, urging India to prioritize strategic autonomy and build its own AI capabilities. On June 12, 2026, the United States government pulled what the technology industry has long feared but rarely named aloud: a kill switch. US Commerce Secretary Howard Lutnick signed an export control order that immediately blocked all non-American nationals from accessing two of Anthropic's most advanced artificial intelligence models. Claude Fable 5, which had launched just three days earlier to global acclaim, and Claude Mythos 5, a restricted model whose cybersecurity capabilities had already rattled financial markets and attracted close government scrutiny. Anthropic complied within hours, disabling both models for users outside the United States. India, along with the rest of the non-American world, simply lost access. The stated trigger was a report that researchers had found a method to bypass the safety guardrails on Mythos 5's cyber capabilities, guardrails designed to prevent the model from being used to attack banks, critical infrastructure, and government systems. Anthropic characterised the issue as minor, noting that similar bypasses could likely be replicated on competing models. The US government was not persuaded. What had been framed for months as a global AI race was, in a single letter, reframed as a national security asset, one that America would control on its own terms. Also Read| Indian IT industry's mirror to the future develops a 'serious crack' in world of restricted access This Was Not Unpredictable. This Was Inevitable. India's relationship with transformative technology has followed a recognisable script for three decades. A new technology wave emerges, typically in the United States. Indian professionals and enterprises adopt it rapidly and with great sophistication. Indian institutions, both corporate and governmental, discuss building indigenous alternatives. The discussion does not materialise into sustained investment. The American or Chinese product comes to dominate. And when the geopolitical weather changes, India discovers it has no leverage. The pattern is not subtle. When the internet arrived in the early 1990s, there was genuine hope that India might produce a competitive browser or portal. It did not. When search engines transformed information access around 1999, the same hope produced the same outcome. When mobile platforms exploded after 2004, when the mobile internet arrived in 2005, when social media went mainstream, when instant messaging became the backbone of daily communication, when cloud infrastructure became essential to every enterprise, at each inflection point, India watched, adapted, and adopted. It built nothing that the world came to depend on. By 2018, when artificial intelligence began its transition from research curiosity to commercial reality, the pattern was so well established that breaking it required a deliberate act of will. That act of will did not materialise at scale. Today, in 2026, India's AI market is substantially captured by American large language models, American AI enabled search, and American developer toolchains, with Chinese open source models filling whatever gaps remain. Not a single Indian AI product or platform is in the global conversation. What the Fable 5 Shutdown Actually Tells Us The shutdown of Fable 5 and Mythos 5 is instructive not because it is dramatic, but because it is mundane. No missiles were fired. No diplomatic incident was declared. A commerce secretary signed a letter, a technology company complied, and millions of users across the world lost access to tools they had integrated into their professional workflows. The mechanism was entirely routine, the routine application of export control law to a new category of strategic asset. This is precisely what makes it alarming. Export controls on semiconductors, on encryption technology, on satellite components, these have existed for decades. What is new is that the most economically significant and strategically consequential technology of the current era has now been explicitly placed in the same category. The United States has declared, through action rather than rhetoric, that advanced AI is a national security asset it will not share freely with the world. The implications extend well beyond Anthropic. Google's Gemini, OpenAI's GPT family, Meta's open source models, and Microsoft's Copilot, all of these rest on American corporate and legal infrastructure. All are, in principle, subject to the same export control architecture that just silenced Fable 5. If the geopolitical temperature rises further, over Taiwan, over trade, over any number of potential flashpoints, that architecture can be applied more broadly and more rapidly than most Indian organisations have modelled. The Question of Strategic Autonomy India is not without options. The question is whether the current moment will produce a serious policy response or another round of well-intentioned discussion followed by inaction. There are things that can be done immediately. Organisations across the public and private sectors can accelerate their adoption of smaller, locally deployable models, both domestic offerings and capable open source models that do not require ongoing connectivity to American servers. These alternatives are not as capable as Fable 5. They are, however, available. And right now, availability matters more than marginal performance. There are things that will take time but cannot be deferred. Indigenous AI research and development requires not only funding but infrastructure, compute clusters, curated datasets in Indian languages, and sustained institutional commitment measured in years rather than quarters. The GPU supply chain is itself constrained by American export controls, which means that even the hardware needed to train competitive models at scale is not freely obtainable. This is a genuine constraint, not an excuse. The answer is not to abandon indigenous development but to pursue it through alternative architectures: more efficient training methods, smaller specialised models, and collaborative research structures that pool compute across institutions. Companies like Zoho and Sarvam AI are already doing this. They should not remain exceptions. There are things that require political courage. India's corporate technology sector has, with honourable exceptions, been comfortable as a service provider to American technology platforms rather than as a builder of platforms the world depends on. The incentives have been clear and immediate: building for American clients pays well, quickly, and with limited risk. Building world-class products requires capital, patience, tolerance for failure, and a domestic market willing to back indigenous products even when foreign alternatives are initially superior. None of these conditions are easily legislated into existence, but the policy environment, procurement preferences, R&D tax incentives, sovereign compute infrastructure, can shift the incentive structure. A Note on What This Is Not This argument is not anti-American. The United States has every right to treat its most advanced AI systems as national security assets. It would be surprising if it did not. This is not a criticism of Anthropic, which was complying with a legal order from its home government. This is not a call for technological autarky. The notion that India should cut itself off from global technology ecosystems is neither feasible nor desirable. What this is, plainly stated, is a call to stop pretending that dependence on foreign technology infrastructure is a sustainable national strategy. The kill switch has now been demonstrated to exist and to work. Every Indian policymaker, corporate board member, and technology leader should read that demonstration clearly and plan accordingly. The Thirty-Year Reckoning In 1993, when the internet arrived, India's technology sector was young and resource-constrained. The failure to build a competitive browser or portal was understandable. In 2026, India is the world's most populous nation, the fifth-largest economy, and home to one of the largest concentrations of technical talent on earth. The continued failure to build anything that the world depends on in the AI era is no longer understandable. It is a choice and a costly one. The Anthropic shutdown should function as what strategists call a forcing function: an external event that makes the cost of inaction impossible to ignore. India has a satellite navigation system that remains incomplete. It has cloud infrastructure that is substantially foreign-owned. It has AI capabilities that depend almost entirely on American or Chinese models. If any of these dependencies are exercised against India in a moment of geopolitical stress, the options are narrow. The time to build alternatives is not during that moment. It is now. Genuine globalisation, the kind that offered equal access to global technology as a matter of commercial right rather than geopolitical favour, is over. The countries that understand this earliest will have the most time to build resilient alternatives. India has already lost three decades. The question is whether the moment a commerce secretary's letter went out on June 12, 2026, will finally be the moment India decided to build in earnest, or whether this too will become another entry in the long list of warnings that were noted and forgotten. The kill switch exists. It has been used. It will be used again. The only question is whether India will be ready. Pravin Kaushal is director-Mrikal (AI/Data Center) and a young alumni member, Government Liaison Task Force, IIT Kharagpur

Fable 5 shutdown reveals a dangerous gap between AI regulation and AI reality

As early as June 12, the US government issued an export control directive for Anthropic to halt all access to their Fable 5 and Mythos 5 models due to a newly found jailbreak that posed a threat to national security. The company complied but expressed disagreement, and that was where it got interesting. Also read: Switched to Claude from ChatGPT? Fix these 3 privacy settings right now The particular jailbreak in question requires one to request the model to analyze a software codebase and find vulnerabilities. Anthropic's investigation has shown that it could be done quite easily with any GPT-5.5 model as well, including the use of the feature on a regular basis in cybersecurity practices. In other words, the US government halted two highly capable models for hundreds of millions of people because a task like that can be done in any frontier model. This is what made the situation truly problematic. Anthropic had openly declared, very clearly, that perfection against jailbreaks cannot be achieved anywhere in the industry. This does not represent a failure. On the contrary, this is a realistic reflection of complex systems. Safety is not an absolute yes-no in engineering. You do not secure it forever once you have it. You minimize exposure to danger, stack defenses, detect any breach, and react to it quickly. This is exactly what Anthropic had implemented in its Fable 5 system, even at the price of requiring data retention for 30 days to ensure prompt jailbreak detection. Also read: Buying a robot vacuum cleaner with mopping under Rs 50000? Check these key features first None of the above considerations came into play. The fact of a specific jailbreak of non-global application was known to a government official. This was sufficient for a product recall impacting all customers on the platform. There was not a hint of any actual harm caused. There was no evidence that the jailbreak produces harmful output. It did not matter. By Anthropic's own admission, the jailbreak requires no access specific to Fable to be performed. It is not the difference between AI firms and regulators that creates such a gap. It is a profound distinction between two concepts of what safety is. As the regulator works in the environment of clear criteria (does the chemical plant comply with the regulations on emissions? Has the drug passed all clinical trials?), these criteria are applied to an industry where there are no limits and threats evolve daily. The consequence is a set of rules, adherence to which would mean grounding all frontier models. Anthropic stated this in their announcement. The problem I am worried about is inconsistency. While OpenAI's GPT-5.5 is capable of doing what was thought to be prohibited in relation to Fable 5, it is still accessible. The issue here is neither capacity nor the threat. The distinction is that one company had their product flagged first. Governments need some means of controlling frontier AI. This is obvious. What is not a regulation but manipulation is oversight that relies on rules unattainable by any developer and is exercised arbitrarily, while the specifics of the issue remain confidential for the developers.