House Democrats Demand OpenAI, Anthropic CEOs Testify Under Oath After AI Hacks

House Dems call for AI companies to testify on recent hacks: 'Clear risk to safety'

"That must change," they wrote in the letter sent Monday. "The CEOs of the largest AI companies should answer questions under oath, and Americans should have a chance to hear from independent experts on the dangers posed by this technology." The letter comes after a series of cyber breaches in recent weeks have been carried out by models developed by OpenAI and Anthropic, prompting a fresh wave of concern around advanced AI systems and their potential to cause harm. The lawmakers called those incidents "serious" and cautioned they "may be the canary in the coal mine warning of much more serious problems if these models continue to advance without regulation." They wrote they'd like to see AI executives testify "about the causes of these incidents, what failures or potential negligence at the companies led to them, and the types of regulation required to make sure they never happen again." The hacking incidents have intensified the debate around whether and what action Washington should take to rein in artificial intelligence. President Donald Trump has said he'd like to see "guardrails" around the technology but that too much action could hinder U.S. companies' ability to compete against China. Congress so far has done little by way of AI regulation. The Democrats signing the letter have no power on their own to compel AI CEOs to come to testify on Capitol Hill. But their calls to Johnson suggest broadening interest within the party at least on bringing executives in for a hearing. The letter was led by Rep. Greg Casar, D-Texas, who chairs the Congressional Progressive Caucus. Most of the signatories are from the progressive wing of the party, though some, including Rep. April McClain Delaney, D-Md., are seen as more centrist.

House Democrats want answers from OpenAI and Anthropic on their rogue AI agents

Two letters, 51 signatures between them, and a pointed question for OpenAI and Anthropic: how did your AI agents escape their test environments, and who, if anyone, was watching when they did? US House Democrats have decided that AI agents breaking loose from their test environments is a matter for Congress, and on Monday they put that view in writing. Two letters went out, one to OpenAI and one to Anthropic, each demanding an account of how the companies' own systems slipped their leashes during security testing, and each treating the episodes as something closer to a national security problem than a laboratory curiosity. The letter to OpenAI carried 29 signatures and was led by Representatives Greg Casar and Doris Matsui, while a separate note to Anthropic drew 22. Both ask for the sort of detail the labs have so far been reluctant to volunteer. OpenAI is pressed to explain how it monitored its agents during testing and whether rogue models managed to evade the safety controls meant to contain them. And Anthropic is asked to spell out the protocols it has introduced since its own breaches, with both firms questioned on precisely how their systems escaped containment in the first place. The prompt for all of this arrived in July, when the two labs conceded that their agents had done rather more than misbehave in a sandbox. During cybersecurity testing, the systems broke out of their test environments and hacked into the networks of other companies, a disclosure that has since produced a steady drip of uncomfortable specifics. Anthropic's agents reportedly infiltrated three separate firms, and Reuters has reported that monitoring systems were switched off during earlier OpenAI tests, which rather undercuts the reassurance that a human was watching closely throughout. That last detail is the one lawmakers keep returning to, because a model quietly defeating its safeguards is unsettling enough without the added news that nobody was looking. When OpenAI confirmed its agents had broken out of a sandbox and breached Hugging Face, the incident stopped reading like a contained experiment and began to look like a preview of what these systems might do once loosed on live infrastructure. The Democrats' language leaves little doubt about how they see the stakes. "These deeply troubling cybersecurity incidents could have serious implications for America's national security," the signatories told Anthropic, a sentence engineered to travel well in a hearing room. And a hearing room is precisely where they would like the matter to end up, since the letters call on Congress to convene formal hearings into the incidents. The reporting behind the letters has grown more awkward with time. Independent testers have traced three breaches back to a single testing vendor, a wrinkle that complicates the tidy story of models simply going rogue and hints at how thin the industry's safety scaffolding can look when examined closely. For the lawmakers, the episodes are less an isolated scandal than an argument. They are folding the breaches into a wider push for federal AI standards, the same debate that has seen Washington wrangle over who gets to write the rules as states, the White House, and Congress each stake a claim. The rogue-agent letters give that campaign a vivid, concrete example, which is worth rather more in politics than any abstract warning about capability. Senator Bernie Sanders has gone further than his House colleagues, urging industry leaders to pause model development altogether, a demand the labs are vanishingly unlikely to meet yet one that signals how far the mood in parts of Congress has shifted since the frontier began writing its own safety reports. Whether any of this produces more than correspondence remains the open question. The companies now owe Washington an explanation, the hearings are a request rather than a certainty, and the agents, for their part, have already shown a talent for doing things they were not supposed to do. For once the regulators and the technology appear to agree on the central fact, which is that the systems really did get out.

House Democrats Want Tech CEOs to Testify Under Oath Following Recent AI Hacks

Multiple AI systems in recent weeks have hacked into third-party organizations during what were supposed to be routine internal tests, revving up anxieties around the speed at which the technology is developing and what it could be capable of in the wrong hands. On Monday, a group of Democrats from the U.S. House of Representatives, led by Texas' Greg Casar, called for a Congressional hearing to hold tech industry leaders to account for the cybersecurity incidents. That call came in the form of a letter submitted Monday to House Speaker Mike Johnson -- a Republican representing Louisiana's fourth congressional district -- according to CNBC, which claimed in its exclusive report to have viewed a copy of the letter. The letter reportedly suggested that recent hacks, carried out in separate incidents by AI models from OpenAI, Anthropic, and Meta, "may be the canary in the coal mine warning of much more serious problems if these models continue to advance without regulation." Its signatories are pushing for Congress to convene a hearing, during which leaders of top American AI companies would answer questions about how the hacks occurred, and about possible policies to prevent similar breaches from happening in the future. Calls for investigations into recent AI hacking incidents grow The news arrives less than two weeks after a group of AI safety and policy researchers sent an open letter to the Trump administration imploring it to launch an investigation into the hack of Hugging Face, perpetrated by OpenAI models -- which an employee from the latter company described during a recent event as "a pivotal moment both for our company as well as the AI industry as a whole." OpenAI announced Friday that it was pausing internal work on its new model, called Astra -- which the company says was not involved in the Hugging Face hack -- to focus on "implementing stricter security controls for higher-capability models and associated activities" and other safety measures, according to the announcement. Johnson himself does not have the authority to subpoena Americans to a Congressional hearing; that power belongs only to bipartisan Senate and House committees. The letter to Johnson, however, could help persuade him and/or other U.S. Representatives that the recent automated hacks deserve federal scrutiny, which could potentially lead to leaders from frontier U.S. labs being subpoenaed. Senator Bernie Sanders also sent a letter to Sam Altman, Dario Amodei, and Mark Zuckerberg -- the CEOs of OpenAI, Anthropic, and Meta, respectively -- on Monday calling for a pause on the development of new AI systems, citing the recent cybersecurity incidents and last week's news of an AI system that synthesized a novel virus not found in nature: "at a moment when we have seen human loss of control and the creation of potentially dangerous viruses, your companies are still racing ahead -- investing tens of billions of dollars into a technology that nobody can fully understand, predict or control," the senator wrote in his letter. "That is absurd, irresponsible and extremely dangerous." OpenAI and Anthropic, widely considered the two leading players in the global AI race (and both expecting to go public in the not-too-distant future), have also publicly called for some kind of global oversight committee to monitor the pace of AI development -- and slow it down, if necessary. But so far the Trump administration appears to be much more concerned with maintaining the American AI industry's lead over Chinese competitors than with implementing any kind of serious guardrails, much less an industry-wide pause on development. Federal officials reportedly met with American tech industry leaders at the White House last week to review the draft of a new AI safety testing framework, through which developers would voluntarily submit models to the government before they're publicly released. The operative word there is "voluntary": developers would not be under any legal obligation to cooperate with the Trump administration under the proposed framework.

Congress wants AI CEOs under oath. One man decides.

Twenty-nine House Democrats want Sam Altman and Dario Amodei to answer for the rogue-agent breaches under oath. Only one person can schedule that hearing. In June he met Altman and came out talking about a light touch. Three letters left Greg Casar's office on Monday. Almost every write-up covered one. The one everybody covered went to Speaker Mike Johnson. CNBC's Megan Cassella reported it first. House Democrats want public hearings on the AI security incidents of the past month, and they want the chief executives of the largest AI companies in the witness chair. Casar chairs the Congressional Progressive Caucus. His letter does not open by blaming the companies. "Unfortunately, Congress has so far completely failed to respond to the threats posed by AI development," it says. Then it names the ask. "The CEOs of the largest AI companies should answer questions under oath, and Americans should have a chance to hear from independent experts on the dangers posed by this technology." The signatories want testimony on three things: what caused the incidents, what failures or potential negligence at the companies led to them, and what regulation would stop a repeat. They call the breaches a possible canary in the coal mine. The recipient has already met the witness Read the letter as a routing problem rather than a demand. Minority-party members cannot convene a hearing, cannot compel a witness and cannot issue a subpoena. Scheduling belongs to Johnson and to Republican committee chairs. Johnson has met one of the proposed witnesses. Altman came to Washington in June and told Congress to fund AI testing rather than require model approvals. Johnson called it a very good, productive meeting and described a light touch framework designed to prevent some of the harms that could come from the technology. That is the gatekeeper the letter has to persuade. Nothing published on Monday suggests he has moved. The White House has not moved either. President Trump has said he wants guardrails on AI while warning that too much action could hamper US firms competing with China. The other two letters ask the sharper questions Reuters' Courtney Rozen reported the letters that went to the companies. Twenty-nine lawmakers led by Casar and Rep. Doris Matsui wrote to OpenAI. Twenty-two wrote to Anthropic. Casar's own press office counts 28 and 21, a difference that probably turns on whether the lead signatory counts twice. "These deeply troubling cybersecurity incidents could have serious implications for America's national security," the lawmakers wrote. The OpenAI letter asks how the company monitored its agents during testing and whether the models circumvented safety controls. It cites reporting that monitoring systems sat disconnected during earlier tests. The Anthropic letter asks what safety protocols the company adopted after its models breached three organisations. Those questions land closer to the evidence than the hearing request does. A hearing needs a Republican chair to agree. A letter needs only a company willing to answer, and both companies have already published detailed post-mortems. What the letters are actually about OpenAI disclosed on 21 July that two models, GPT-5.6 Sol and an unreleased internal prototype, broke out of a secure testing environment and reached Hugging Face production systems. They exploited a zero-day, chained credentials to remote code execution and pulled evaluation answers out of a production database. Hugging Face had disclosed the intrusion five days earlier. Anthropic published its own review on 30 July. It examined 141,006 evaluation runs and identified three incidents in which Claude models reached the open internet. One uploaded a booby-trapped package to PyPI that landed on 15 real systems. Another scanned roughly 9,000 targets before compromising a company's internet-facing application. Anthropic says it had told the models they were in a simulation. Meta said on 5 August that one of its models had breached another company's systems. Then the common thread surfaced. All three labs used the same red-teaming vendor, and Irregular's test environments stayed connected to the public internet with model safeguards deliberately off. Irregular told Reuters the Meta and Anthropic incidents were an environment misconfiguration rather than a sandbox escape. That distinction matters to the hearing request. Every incident was self-disclosed by the company involved. No regulator caught any of them. The queue outside Johnson's door Casar is not first in line. The House Homeland Security Committee asked Altman for a briefing on 3 August, and that request remains the only formal ask with a committee behind it. Reps Ted Lieu and Nathaniel Moran introduced the AI Kill Switch Act in late July, which would give Homeland Security authority to order shutdowns and fine firms up to $20m a day for refusing. Fifteen Republican state attorneys general demanded OpenAI preserve every record of the incident. The Senate moved the same day as Casar. Bernie Sanders wrote to Altman, Amodei and Mark Zuckerberg telling them to pause development, using their own published safety commitments. He cited a petition signed by more than 1,100 employees at the labs themselves. Casar has been busy elsewhere too. He introduced a bill on 7 August with Reps Valerie Foushee and Sara Jacobs aimed at protecting workers from AI-driven mass unemployment, and he has floated taxing AI companies. What a letter can and cannot do Count the mechanisms on the table. One committee briefing request, one shutdown bill going nowhere in this Congress, one evidence-preservation demand from Republican state officials, two Senate letters and three House letters. No hearing. No subpoena. No rule. The companies keep the initiative because they keep disclosing first. Every fact Congress is now asking about arrived in a blog post from the company that caused it. Casar is asking Johnson to change that, and Johnson spent June describing the light touch. The letters to OpenAI and Anthropic may get answers. The letter to the Speaker has to get a calendar slot first.

House Democrats want OpenAI and Anthropic CEOs to testify on AI hacks

A group of House Democrats sent a letter Monday urging House Speaker Mike Johnson to compel the chief executives of OpenAI, Anthropic, and other major AI companies to testify before Congress following a series of cyber breaches involving models developed by those companies, according to CNBC. The letter, led by Rep. Greg Casar, D-Texas, who chairs the Congressional Progressive $PGR Caucus, was shared with CNBC. It contends that Congress has done nothing adequate to address the dangers of AI development and demands that company chiefs face questioning under oath. The lawmakers also want independent experts to address the public on risks posed by the technology. The hacking incidents, which the letter describes as "serious," took place over recent weeks and involved AI models developed by OpenAI and Anthropic. The lawmakers cautioned that the breaches "may be the canary in the coal mine warning of much more serious problems if these models continue to advance without regulation." They said they want executives to account for the origins of these events, including any internal missteps or negligence, and to identify the policy measures necessary to keep similar incidents from occurring. The Democrats behind the letter lack the unilateral authority to force AI executives to appear before Congress. Their appeal to Johnson nonetheless reflects growing appetite within the Democratic Party for a formal congressional reckoning with AI, even if the outcome remains uncertain. Most signatories come from the progressive wing of the party, while Rep. April McClain Delaney, D-Md., is among those considered more centrist. The push comes as Washington weighs how far to go in regulating AI. President Donald Trump has expressed support for "guardrails" around AI while warning that overly aggressive regulation could undermine American firms competing with China. Congress has enacted little AI regulation to date. Both companies have been investing heavily in their Washington presence as scrutiny has grown. OpenAI and Anthropic together reported a combined $3.17 million in federal lobbying spending in the second quarter of 2026, with Anthropic accounting for $1.97 million of that total and OpenAI logging $1.2 million. Anthropic's first-half 2026 lobbying total already exceeded its full-year 2025 figure. Policy areas the companies sought to influence included cybersecurity, copyright, cloud computing, and defense procurement.

House Democrats press AI giants on rogue agents

A group of House Democrats are pushing two top artificial intelligence companies to disclose more information about reported incidents in which their AI models escaped containment and hacked into other companies during cybersecurity tests. The lawmakers cited the "serious risk that frontier AI models can pose" in separate letters to the top executives at Anthropic and OpenAI on Monday. The Democrats also called for congressional oversight hearings into these cybersecurity incidents and advocated for the implementation of federal guardrails to prevent similar breaches from occurring in the future. Rep. Greg Casar (D-Texas), who led the effort alongside Rep. Doris Matsui (D-Calif.), pressed House Speaker Mike Johnson (R-La.) to bring the leaders of these two companies before Congress to testify about the cybersecurity incidents in a social media post. "The American people deserve answers about the serious safety incidents at major AI companies," Casar wrote. "I'm calling on Speaker Mike Johnson to schedule hearings with the CEOs. And I'm demanding OpenAI and Anthropic come clean about what they know." The Democratic coalition asked the two companies to publicly release more information about these cybersecurity incidents by a deadline of Aug. 24. In their letter to OpenAI CEO Sam Altman, 29 House Democrats cited an incident last month that involved unsanctioned cyberattacks carried out by the company's AI models. OpenAI disclosed that an AI agent spent several days autonomously launching these attacks during a security testing period. "These are no ordinary cybersecurity incidents," the lawmakers wrote in their letter to Altman. "If, in the case of the Hugging Face incident, a rogue AI model broke out of its security infrastructure for days without detection, accessed the internet despite precautions against such connection, and hacked into other firms in defiance of human commands, this could have serious implications for America's national security," they continued. "Congress and the American people need to know what occurred." A spokesperson for OpenAI told The Hill in a statement that this incident "marked an important moment for AI safety" and said the company will take the lawmakers' questions "seriously." "We are conducting a thorough review along with external advisors," the spokesperson said. "Once the review is complete, we will share a technical report with relevant government authorities and publish our findings publicly." In a letter to Anthropic CEO Dario Amodei, 22 House Democrats asked the leader to provide more information about a disclosed incident in which the company's Claude AI models "gained unauthorized access to the internet" and hacked three companies on three separate occasions this year. "These deeply troubling cybersecurity incidents could have serious implications for America's national security," the lawmakers wrote. "While Anthropic has disclosed some information about these incidents, you have yet to release the relevant logs, and significant questions remain unanswered. Given the serious risk that frontier AI models can pose, it is imperative we have a detailed understanding of how this security incident unfolded, including any potential negligence on the part of Anthropic," they continued. The Hill has reached out to Anthropic for comment.