House Dems call for AI companies to testify on recent hacks: ‘Clear risk to safety’
A group of House Democrats is calling on leaders of Anthropic, OpenAI and other AI companies to testify in Congress about recent hacking incidents.
A congressional summons, even an informal one, is how Washington tells an industry that its honeymoon is over. The call for Anthropic and OpenAI to explain recent hacks is less about the incidents themselves and more about establishing the premise that these companies hold systemic risk.
Notice the framing: “clear risk to safety.” That is not a technical assessment; it is a legal and political predicate. Once a witness agrees to that framing under oath, the debate is no longer about whether AI should be regulated, but how.
The choice of companies matters. House Democrats did not reach for social media platforms or software vendors. They targeted the two labs that define the frontier of the technology. That tells you where the credible threat is thought to live.
For Anthropic and OpenAI, testifying is a double-edged opportunity. They get to project responsibility and offer a narrative of vigilance. But they also invite scrutiny of their security practices, their training data, and their deployment decisions. In a public hearing, every question has a hidden agenda.
What this is not yet is legislation. But hearings are how the legislative machinery begins to move. The record they produce becomes the justification for future rules, for budget decisions, for oversight mandates. The actual hacking incidents will fade; the testimony will endure.
There is also a quiet political calculation at work. By focusing on national security and public safety, House Democrats are choosing a frame that is harder to wave away as anti-innovation. They are betting that the AI industry would rather cooperate than be cast as the reckless party in a breach-pocked landscape.
The real question now is how the companies respond. Whether they send their CEOs or their deputies, whether they arrive with technical detail or diplomatic language, will tell observers just how scared these firms really are. The subpoena is not yet a subpoena, but the direction is clear.