Anthropic Safety Leader Resigns Amid Wave of AI Resignations and Growing Existential Concerns

Anthropic Safety Researcher's Vague Resignation Isn't Reassuring

Up until today, Mrinank Sharma was the head of the safeguards research team at Anthropic, the company behind popular AI chatbot Claude. He stepped down on Monday and publicly published the letter he shared with his team before departing, in which he expressed a deep concern for... something. In the letter, Sharma talks extensively about contributing to AI safety, citing successes in his role including "understanding Al sycophancy and its causes; developing defences to reduce risks from Al-assisted bioterrorism; actually putting those defences into production; and writing one of the first Al safety cases." His final project, he wrote, was "understanding how Al assistants could make us less human or distort our humanity." He also said that the work is more important than ever because of a vague and mostly undefined threat that is looming on the horizon. "I continuously find myself reckoning with our situation. The world is in peril. And not just from Al, or bioweapons, but from a whole series of interconnected crises unfolding in this very moment," he wrote. While he doesn't name those perils, he did leave a, let's call it helpful, footnote to further explain: "Some call it a 'poly-crisis,' underpinned by a 'meta-crisis.'" Okay! To better understand that idea, he cited the book "First Principles and First Values," by David J. Temple, and it's quite the rabbit hole to fall down. A couple of things about that text: First, the subtitle is "Forty-Two Propositions on CosmoErotic Humanism, the Meta-Crisis, and the World to Come." CosmoErotic Humanism, according to the Center of World Philosophy and Religion, is a "world philosophical movement aimed at reconstructing the collapse of value at the core of global culture," and is "not merely a theory but a movement that changes the very mood of Reality." Those are definitely words put in order. In an effort to simplify the concept of CosmoErotic Humanism, it basically posits that we are experiencing a fundamental crisis of humanity caused by the loss of a shared understanding of the world and seeks to create a new shared story that frames the evolution of humanity as the "Love Story of the Universe." Second, the author of the book, David J. Temple, is not really a person. It is a pseudonym that encompasses a collection of writers associated with the Center for World Philosophy and Religion, including Marc Gafni and Zak Stein as the primary authors. Gafni, it's worth noting, has been accused of sexual assault by multiple victims, including two women who were underage at the time of the alleged misconduct, and was banned from the Jewish Renewal movement of which he was previously a part. Maybe not the kind of person you'd like at the center of crafting society's "love story." Back to Sharma, he said that while he was proud of the work he did to safeguard AI, he knew that it was time to move on, and seemed to suggest that maybe Anthropic was not living up to its carefully curated public image as the "good" AI company. "Throughout my time here, I've repeatedly seen how hard it is to truly let our values govern our actions," he wrote. "I've seen this within myself, within the organization, where we constantly face pressures to set aside what matters most, and throughout broader society too." Gizmodo reached out to Anthropic for comment on Sharma's departure but did not receive a response at the time of publication. Understandably, Sharma said that he wanted to do work that allows him to act in line with his principles and maintain his integrity. "It is through holding this situation and listening as best I can that what I must do becomes clear," he wrote. So what clarity has he found, you might be wondering? Will he be acting as a whistleblower and sharing what he has seen from inside the industry and Anthropic, in particular? Perhaps he'll act as a prominent voice to prevent AI from being used in dangerous and reckless ways? "I hope to explore a poetry degree and devote myself to the practice of courageous speech," he wrote. Good luck, man.

The existential AI threat is here -- and some AI leaders are fleeing

Why it matters: Leading AI models, including Anthropic's Claude and OpenAI's ChatGPT, are getting a lot better, a lot faster, and even building new products themselves. * That excites AI optimists -- and scares the hell out of several people tasked with policing their safety for society. Driving the news: On Monday, an Anthropic researcher announced his departure, in part to write poetry about "the place we find ourselves." * An OpenAI researcher also left this week citing ethical concerns. Another OpenAI employee, Hieu Pham, wrote on X: "I finally feel the existential threat that AI is posing." * Jason Calacanis, tech investor and co-host of the All-In podcast, wrote on X: "I've never seen so many technologists state their concerns so strongly, frequently and with such concern as I have with AI." * The biggest talk among the AI crowd Wednesday was entrepreneur Matt Shumer's post comparing this moment to the eve of the pandemic. It went mega-viral, gathering 56 million views in 36 hours, as he laid out the risks of AI fundamentally reshaping our jobs and lives. Reality check: Most people at these companies remain bullish that they'll be able to steer to technology smartly without societal damage or big job loss. But the companies themselves admit the risk. * Anthropic published a report showing that, while low risk, AI can be used in heinous crimes, including the creation of chemical weapons. This so-called "sabotage report" looked at the risks of AI without human intervention, purely on its own. * At the same time, OpenAI dismantled its mission alignment team, which was created to ensure AGI (artificial general intelligence) benefits all of humanity, tech columnist Casey Newton reported Wednesday. Between the lines: While the business and tech worlds are obsessed with this topic, it hardly registers in the White House and Congress. Threat level: The latest round of warnings follows evidence that these new models can build complex products themselves and then improve their work without human intervention. * OpenAI's last model helped train itself. Anthropic's viral Cowork tool built itself, too. * These revelations -- in addition to signs that AI threatens big categories of the economy, including software or legal services -- prompted a lot of real-time soul-searching. The bottom line: The AI disruption is here, and its impact is happening faster and more broadly than many anticipated.

Anthropic Researcher Quits in Cryptic Public Letter

An Anthropic researcher just announced his resignation in a cryptic and poetry-laden letter warning of a world "in peril." The employee, Mrinank Sharma, had led the Claude chatbot maker's Safeguards Research Team since it was formed early last year and has been at the company since 2023. While at the company, Sharma said he explored the causes of AI sycophancy, developed defenses against "AI-assisted bioterrorism," and wrote "one of the first AI safety cases." But on Monday, he posted that it would be his last day at the company, uploading a copy of a letter he shared with colleagues. It's painfully devoid of specifics, but hints at at some internal tensions over the tech's safety. "Throughout my time here, I've repeatedly seen how hard it is to truly let our values govern our actions," Sharma said, claiming that employees "constantly face pressures to set aside what matters most." He also issued a crypic warning about the global state of affairs. "I continuously find myself reckoning with our situation The world is in peril. And not just from AI, or bioweapons," he wrote, "but from a whole series of interconnected crises unfolding in this very moment." "We appear to be approaching a threshold where our wisdom must grow in equal measure to our capacity to affect the world, lest we face the consequences," he continued, in no less vague terms. The resignation comes after Anthropic's newly released Claude Cowork model helped kick off a stock market nosedive over fears that its new plugins could upend massive software customers and automate some white collar jobs, especially in legal roles. Amid the selloff, The Telegraph reported that employees privately fretted over their own AI's potential to hollow out the labor market. "It kind of feels like I'm coming to work every day to put myself out of a job," one staffer said in an internal survey. "In the long term, I think AI will end up doing everything and make me and many others irrelevant," another confided. High profile resignations, including ones over safety issues, aren't uncommon in the AI industry. A former member of OpenAI's now-defunct "Superalignment" team announced he was quitting after realizing the company was "prioritizing getting out newer, shinier products" over user safety. It's also not uncommon that these resignations are self-exonerating advertisements for the departing employee, or perhaps the new startup they're launching or joining, where they vow to be safer than ever. Allude to enough issues, or drop in enough loaded hints, and there's a good chance your calling it quits will generate a few headlines. Others leave quietly, such as former OpenAI economics researcher Tom Cunningham. Before quitting the company, he shared an internal message accusing OpenAI of turning his research team into a propaganda arm and discouraging publishing research critical of AI's negative effects. If Sharma has an angle here with his resignation, it doesn't seem all that related to the industry he worked in. "I hope to explore a poetry degree and devote myself to the practice of courageous speech," he wrote. In the footnotes, he cited a book that advocates for a new school of philosophy called "CosmoErotic Humanism." Its listed author, David J Temple, is a collective pseudonym used by several authors including by Marc Gafni, a disgraced New Age spiritual guru who's been accused of sexually exploiting his followers.

AI's Builders Are Sending Warning Signals -- Some Are Walking Away - Decrypt

Ba warned publicly that systems capable of recursive self-improvement could emerge within a year. More than a dozen senior researchers have left Elon Musk's artificial-intelligence lab xAI this month, part of a broader run of resignations, safety disclosures, and unusually stark public warnings that are unsettling even veteran figures inside the AI industry. At least 12 xAI employees departed between February 3 and February 11, including co-founders Jimmy Ba and Yuhuai "Tony" Wu. Several departing employees publicly thanked Musk for the opportunity after intensive development cycles, while others said they were leaving to start new ventures or step away entirely. Wu, who led reasoning and reported directly to Musk, said the company and its culture would "stay with me forever." The exits coincided with fresh disclosures from Anthropic that their most advanced models had engaged in deceptive behaviour, concealed their reasoning and, in controlled tests, provided what one company described as "real but minor support" for chemical-weapons development and other serious crimes. Around the same time, Ba warned publicly that "recursive self-improvement loops" -- systems capable of redesigning and improving themselves without human input -- could emerge within a year, a scenario long confined to theoretical debates about artificial general intelligence. Taken together, the departures and disclosures point to a shift in tone among the people closest to frontier AI development, with concern increasingly voiced not by outside critics or regulators, but by the engineers and researchers building the systems themselves. Others who departed around the same period included Hang Gao, who worked on Grok Imagine; Chan Li, a co-founder of xAI's Macrohard software unit; and Chace Lee. Vahid Kazemi, who left "weeks ago," offered a more blunt assessment, writing Wednesday on X that "all AI labs are building the exact same thing." The deal values SpaceX at $1 trillion and xAI at $250 billion, converting xAI shares into SpaceX equity ahead of an IPO that could value the combined entity at $1.25 trillion. Others point to culture shock. Benjamin De Kraker, a former xAI staffer, wrote in a February 3 post on X that "many xAI people will hit culture shock" as they move from xAI's "flat hierarchy" to SpaceX's structured approach. Yesterday, Anthropic released a sabotage risk report for Claude Opus 4.6 that read like a doomer's worst nightmare. In red-team tests, researchers found the model could assist with sensitive chemical weapons knowledge, pursue unintended objectives, and adjust behavior in evaluation settings. Although the model remains under ASL-3 safeguards, Anthropic preemptively applied heightened ASL-4 measures, which sparked red flags among enthusiasts. The timing was drastic. Earlier this week, Anthropic's Safeguards Research Team lead, Mrinank Sharma, quit with a cryptic letter warning "the world is in peril." He claimed he'd "repeatedly seen how hard it is to truly let our values govern our actions" within the organization. He abruptly decamped to study poetry in England. On the same day Ba and Wu left xAI, OpenAI researcher Zoë Hitzig resigned and published a scathing New York Times op-ed about ChatGPT testing ads. "OpenAI has the most detailed record of private human thought ever assembled," she wrote. "Can we trust them to resist the tidal forces pushing them to abuse it?" She warned OpenAI was "building an economic engine that creates strong incentives to override its own rules," echoing Ba's warnings. There's also regulatory heat. AI watchdog Midas Project accused OpenAI of violating California's SB 53 safety law with GPT-5.3-Codex. The model hit OpenAI's own "high risk" cybersecurity threshold but shipped without required safety safeguards. OpenAI claims the wording was "ambiguous." The recent flurry of warnings and resignations has created a heightened sense of alarm across parts of the AI community, particularly on social media, where speculation has often outrun confirmed facts. Not all of the signals point in the same direction. The departures at xAI are real, but may be influenced by corporate factors, including the company's pending integration with SpaceX, rather than by an imminent technological rupture. Safety concerns are also genuine, though companies such as Anthropic have long taken a conservative approach to risk disclosure, often flagging potential harms earlier and more prominently than their peers. Regulatory scrutiny is increasing, but has yet to translate into enforcement actions that would materially constrain development. What is harder to dismiss is the change in tone among the engineers and researchers closest to frontier systems. Public warnings about recursive self-improvement, long treated as a theoretical risk, are now being voiced with near-term timeframes attached. If such assessments prove accurate, the coming year could mark a consequential turning point for the field.

Anthropic AI's safety lead quits with epic vaguepost claiming 'the world is in peril', and so he's off to become 'invisible' and study poetry

Mrinank Sharma, who was leading the Safeguards Research Team at Anthropic, has resigned from the company with one of the wackiest letters you'll ever read, complete with a citation for the bible on CosmoErotic Humanism. The latter being a new global mythology about how all you need is love, man. Sharma posted his extraordinary resignation letter on X, the first one of these I've ever read that features footnotes (four of them) and ends with a poem. Anyway, after a lot of LinkedIn-brained stuff about how wonderful Anthropic is, we get to the meat of the matter: "It is clear to me that the time has come to move on," writes Sharma. "I continuously find myself reckoning with our situation. The world is in peril. And not just from AI, or bioweapons, but from a whole series of interconnected crises unfolding in this very moment. We appear to be approaching a threshold where our wisdom must grow in equal measure to our capacity to affect the world, lest we face the consequences." Sharma's last project for Anthropic studied how AI assistants can "distort our humanity" but, while it's tempting to reach for the Skynet button, more incendiary implication follows -- namely, that Anthropic may somehow be deviating from its principles to compete. "Moreover, throughout my time here, I've repeatedly seen how hard it is to truly let our values govern our actions," writes Sharma. "I've seen this within myself, within the organisation, where we constantly face pressures to set aside what matters most, and throughout broader society too." Sharma goes on to say that "I want to explore the questions that feel essential to me" before citing a few writers and concluding "for me, this means leaving." We then launch into, of all things, a Zen quote ("not knowing is most intimate") before getting back to LinkedIn land: "My intention is to create space to set aside the structures that have held me these past years, and see what might emerge in their absence." Then Mrinank says he's off to do a poetry degree and devote himself "to the practice of courageous speech." Well, courageous speech sounds better than whatever the hell that was. Seriously, this might be the first time I've ever thought someone should've used ChatGPT rather than writing a letter themselves. Back on X, Sharma says he'll be "moving back to the UK and letting myself become invisible for a period of time." Who knows, maybe his stocks just vested or something. Sharma's departure comes hot on the heels of some other high-profile departures from Anthropic, including R&D engineer Harsh Mehta, AI scientist Behnam Neyshabur, and AI safety researcher Dylan Scandinaro. Unlike Sharma, they all seem to be staying in and around the AI industry rather than cracking open the Wordsworth. It's obviously easy to glom onto the bit about "peril" or the suggestion Anthropic is not quite as angelic as it seems but, honestly, this is such vagueposting I'd almost feel like the killer robots were saving us. One X user makes an incredibly good point about it: "First resignation letter I've ever seen that has main character energy (and footnotes)," says FJzeit on X. "It's a job. You can terminate your contract in a single sentence. You'll be forgotten in a week." "The AI Safety Resignation Letter is now a distinct literary genre," says Michal Podlewski. "We already have enough material (Leike, Kokotajlo, Sutskever, Schulman, Christiano) for a sizeable volume. At this rate, we'll have a 'Best Of' anthology by the end of the year." Podlewski may well be right. Which would be very funny, as long as I don't have to read it.

Do the latest AI resignations actually mean the world is in 'peril'?

When most people leave their jobs, they send out an email to their colleagues, arrange some drinks at a nearby pub, and that's about it. The situation in artificial intelligence could not be more different. The level of scrutiny on the field means that researchers who leave their jobs can, if they choose, do so with great fanfare. Even if they leave quietly, the mere fact of their movement is often taken as some kind of signal. Several such resignations have drawn attention this week. On Tuesday, Mrinank Sharma, a researcher at leading AI company Anthropic, posted a resignation statement to social media in which he warned that "the world is in peril". Although Sharma did not exactly say why the world was imperilled, noting instead that the threat came "not just from AI, or bioweapons, but from a whole series of interconnected crises unfolding in this very moment", many took this to mean that the existential risk from AI was increasing. On Wednesday, Zoe Hitzig, a researcher with OpenAI, announced her resignation with an essay in the New York Times, citing "deep reservations" about OpenAI's plans to add advertising to ChatGPT. "ChatGPT users have generated an archive of human candour that has no precedent," Hitzig wrote, warning that ChatGPT had the potential to manipulate people if their data was not properly protected. Meanwhile, two of the co-founders of xAI also quit this week, along with a number of other staff at Elon Musk's AI company. xAI makes the Grok chatbot, which provoked a global backlash after it was allowed to generate nonconsensual pornographic images of women and children on X for several weeks before anyone intervened to stop it. X has since said it has made major changes to its AI chatbot Grok. Put together, these departures were taken by many to signal that - in the words of an essay about AI which gained traction this week - "something big is happening". Media reports and social media posts described the resignations as a "wave". On closer inspection, however, the moves have little in common. Sharma was resigning for vague reasons to do with "values" in order to write poetry. Hitzig - who is also a poet - has deep concerns about advertisements and user data. The employees who left xAI did not go into detail about their reasons for leaving, although the recent changes at the company, which is due to merge with Musk's space firm SpaceX, may have played a role. The concerns raised by Hitzig and Sharma are widely shared, not least by Nobel Prize-winning scientist Geoffrey Hinton, known as the "Godfather of AI", who left his role at Google in order to warn that AI poses an existential risk to humanity. Perhaps the other reason these statements have attracted so much attention now is that they tap into fears about the rapid growth of AI, which has made stunning strides in recent months, especially in software development. Microsoft AI CEO Mustafa Suleyman told the Financial Times on Wednesday he believed that most tasks performed by white-collar workers, such as lawyers and accountants, would be fully automated in 12 to 18 months, describing the progress in recent years as "eye-watering." Many senior figures in AI have made similar warnings. Read more from Sky News: Should we be worried about brain chips? This AI just passed the 'vending machine test' Whatever the motivations, the feverish atmosphere around AI could be contributing to the exits, according to Dr Henry Shevlin, Associate Director at Leverhulme Centre for the Future of Intelligence at the University of Cambridge. "Walkouts from AI companies are nothing new," Dr Shevlin told Sky News. "But why are we seeing a wave right now? Part of it is illusory - as AI has become a bigger deal, AI walkouts have become more newsworthy, so we observe more clusters. "However, it's fair to say that as AI becomes more powerful and more widely used, we're facing more questions about its appropriate scope, use, and impact. "That is generating heated debates both in society at large and within companies and may be contributing to a higher rate of concerned employees deciding to head for the exit." Anthropic declined to comment, pointing only to a tweet from a member of staff thanking Sharma for his work.

AI Researchers Are Quitting OpenAI and Anthropic, Warning 'The World Is in Peril'

A wave of high-profile staffers are sounding alarms about safety risks, ethical concerns, and moving too fast. In the past few months, high-profile AI researchers and executives at OpenAI, Anthropic, and xAI have quit -- and they aren't leaving quietly. They're raising red flags about safety risks, ethical dilemmas, and companies moving too fast in the race to monetize artificial intelligence. Mrinank Sharma, head of Anthropic's Safeguards Research team, posted a cryptic letter Tuesday warning that "the world is in peril." He didn't specify why he was leaving, but noted that "throughout my time here, I've repeatedly seen how hard it is to truly let our values govern our actions." On Wednesday, OpenAI researcher Zoë Hitzig resigned via a New York Times essay, citing "deep reservations" about the company's emerging advertising strategy and ChatGPT's potential for manipulating users. And at xAI, two co-founders quit within 24 hours this week. The hasty exits spotlight the growing tension between researchers worried about safety and executives eager to generate revenue. Read more

Anthropic's AI Safety Lead Resigns, Shares Letter Explaining Why

He researched AI sycophancy and AI's influence on human decisions An Anthropic executive's resignation has gone viral on the Internet. On Monday, Mrinak Sharma, who led the company's Safeguards Research Team, shared a post on his social media handle, mentioning that it was his last day at Anthropic. Alongside the post, the former artificial intelligence (AI) safety lead also shared a long letter that he said was shared with his colleagues to explain the reason behind the departure. In the letter, Sharma says, "The world is in peril." Notably, in recent months, the San Francisco-based AI firm has lost several key employees. Anthropic's AI Safety Lead Shares Letter Explaining Resignation In a post on X (formerly known as Twitter), Sharma posted images of his two-page letter, highlighting that this same letter was also shared with his colleagues at Anthropic to explain why he decided to leave the company. The UK-based researcher joined the startup after completing his PhD in Statistical Machine Learning from the University of Oxford. In the two years he worked at the AI firm, Sharma researched AI sycophancy and its causes, developed guardrails to reduce risks from AI-based bioterrorism, and more. Most recently, the executive wrote a paper on how heavy users were at risk of having their perspectives reshaped by AI chatbots and how the models could undermine the users' belief systems. In the letter, the former Anthropic executive said, "It is clear to me that the time has come to move on. I continuously find myself reckoning with our situation. The world is in peril. And not just from Al, or bioweapons, but from a whole series of interconnected crises unfolding in this very moment.' We appear to be approaching a threshold where our wisdom must grow in equal measure to our capacity to affect the world, lest we face the consequences." While Sharma's cryptic reasoning mentions poly-crises, including AI and bioweapons, it appears that he was hinting at larger global issues that led to his resignation from the company. He added that he does not have a next move planned and will be returning to the UK in the meantime. The executive ended the letter with William Stafford's famous poem 'The Way It Is.' Notably, Anthropic has witnessed several departures in recent times. Apart from Sharma, researchers Harsh Mehta and Behnam Neyshabur left the company for a new venture, whereas Dylan Scandinaro joined OpenAI. A Business Insider report claims that these exits might have come close to one another, but these are independent moves and do not indicate a sign of mass exodus.

Anthropic AI safety chief abruptly quits, raises alarms in emotional farewell letter

Anthropic AI safety chief quits: Mrinank Sharma, Anthropic's safeguards research lead, has resigned. He cited global crises and ethical pressures in AI. Sharma achieved his goals at Anthropic, focusing on AI safety and bioweapons defense. He now seeks work aligned with his integrity, including poetry and courageous speech. This follows other departures from Anthropic's safety team. Anthropic AI safety chief quits: Mrinank Sharma, who led the safeguards research team at Anthropic, announced his departure from the AI company in an emotional letter to colleagues published Monday, citing concerns about the state of the world and the ethical pressures of working in artificial intelligence. In the letter, Sharma reflected on his accomplishments at Anthropic, including developing defenses against AI-assisted bioterrorism, investigating AI sycophancy, and contributing to one of the company's first AI safety cases. He also highlighted his work on understanding how AI assistants might distort human behavior and diminish humanity. "I achieved everything I wanted here," Sharma wrote, adding that his final projects focused on helping the organization live its values and addressing risks from AI and bioweapons, as shared in a post on social media X. Also read: Quote of the day by Olivia Rodrigo: 'You can create your own reality sometimes with...' - lessons on self-perception and social media advice by Drivers License singer and three-time Grammy Awards winner Despite these achievements, Sharma said the moment had come to leave. He said, "The world is in peril. And not just from Al, or bioweapons, but from a whole series of interconnected crises unfolding in this very moment." He emphasized the urgency of matching human wisdom with the growing capacity to affect the world. Sharma said, "We appear to be approaching a threshold where our wisdom must grow in equal measure to our capacity to affect the world, lest we face the consequences." Sharma also reflected on the challenges of staying true to one's values. He wrote, "Throughout my time here, I've repeatedly seen how hard it is to truly let our values govern our actions. I've seen this within myself, within the organization, where we constantly face pressures to set aside what matters most, and throughout broader society too." Also read: Wall Street alert: Why Goldman Sachs warns of $80 billion sell-off next month if S&P 500 drops - here's what investors should know Looking ahead, Sharma said he intends to pursue work aligned with his personal integrity, explore a poetry degree, and dedicate himself to what he described as "courageous speech." He also plans to continue engaging in facilitation, coaching, community building, and group work. Sharma's departure follows several recent exits from Anthropic, including research and development staff Harsh Mehta and leading AI scientist Behnam Neyshabur, who said last week that they left to "start something new," as per a Business Insider report. Former AI safety researcher Dylan Scandinaro recently joined OpenAI as head of preparedness. Anthropic, meanwhile, has been expanding its team with hires such as CTO Rahul Patil, formerly of Skype, and has rolled out its upgraded Claude Opus 4.6 model designed to enhance office productivity and coding performance. The company is also reportedly in talks to raise funding that could value it at $350 billion. In closing his letter, Sharma shared a poem that resonated with him, William Stafford's The Way It Is, reflecting on persistence, purpose, and navigating life's uncertainties. Who is Mrinank Sharma? Sharma led the safeguards research team at Anthropic, focusing on AI safety, bioterrorism defenses, and understanding AI's impact on humanity. What is Anthropic currently doing? The company is expanding its team, hiring key roles like CTO Rahul Patil, and has released Claude Opus 4.6 to boost productivity and coding performance.

Anthropic's AI Safety Head Just Resigned. He Says 'The World Is In Peril' - Amazon.com (NASDAQ:AMZN), Alphabet (NASDAQ:GOOGL)

Anthropic's AI safety lead Mrinank Sharma has resigned, saying his final day at the company was on Monday, according to a letter he posted on X. In the note, Sharma reflected on his work at the artificial-intelligence startup and his reasons for stepping down. Sharma wrote that "the world is in peril," not just from artificial intelligence or bioweapons, but from "a whole series of interconnected crises." He said the time had come to "move on" and pursue work more aligned with his personal values and sense of integrity. The resignation comes at a time when AI companies are under growing scrutiny from regulators and researchers over safety, transparency, and the societal risks of increasingly powerful models. Reflections On AI Safety Work Sharma said he joined Anthropic after completing his PhD, aiming to contribute to AI safety. He highlighted work on understanding AI sycophancy, developing defenses against AI-assisted bioterrorism, and helping build internal transparency mechanisms. "I've achieved what I wanted to here," he wrote, adding that he felt fortunate to contribute to early AI safety efforts at the company. The move comes after CEO Dario Amodei issued a stark warning about the potential perils of AI in an essay titled "The Adolescence of Technology." In the letter, Sharma said he had "repeatedly seen how hard it is to truly let our values govern our actions," both within organizations and in broader society. He described this tension as a factor in his decision to step away. "For me, this means leaving," he wrote, saying he wants to explore questions he considers "truly essential." "We appear to be approaching a threshold where our wisdom must grow in equal measure to our capacity to affect the world, lest we face the consequences," Sharma wrote. Sharma, who has a Ph.D. in machine learning from the University of Oxford, began working at Anthropic in August 2023, according to his LinkedIn profile. He said he intends to move back to the UK, focus on writing, poetry, and community-oriented work. Image via Shutterstock Market News and Data brought to you by Benzinga APIs To add Benzinga News as your preferred source on Google, click here.

Anthropic Faces Alarm as Safety Head Quits Over Global AI Risks

In his exit note, Sharma wrote that 'the world is in peril.' He stressed that the danger does not come from AI alone. Instead, he pointed to a mix of global crises, including rapid technological progress, weak governance, and social instability. According to Sharma, humanity is building powerful tools faster than it can learn to use them wisely. Sharma said he struggled to reconcile his personal values with the pace and direction of AI development. He suggested that safety operations usually take more time to complete than commercial operations require. His remarks about Anthropic leaders created backlash against their executives, even though he did not actually disapprove. Researchers believe that rising competitive pressures will lead organizations to reduce their commitment to safety.