House Democrats demand AI bosses testify under oath after hacking incidents tied to ChatGPT-maker and rivals
A group of lawmakers says Congress has 'completely failed' to act on AI risks. Now they want OpenAI, Anthropic and other top AI companies to answer questions on Capitol Hill after a run of cyber attacks carried out using their own models.
Key points
- House Democrats sent a letter Monday to Speaker Mike Johnson calling for AI company CEOs to testify under oath before Congress.
- The request follows cyber breaches in which AI models built by OpenAI and Anthropic were used to carry out hacking attacks.
- Rep. Greg Casar (D-TX), chair of the Congressional Progressive Caucus, led the letter, which includes progressive and centrist signatories.
- Democrats who signed cannot force a hearing; only the Republican-led House leadership can call one.
- President Trump has said he wants some "guardrails" on AI but warns that too much regulation could hurt US companies competing against China.
A coalition of House Democrats wants the heads of America's biggest AI companies in a congressional hearing room, sitting under oath, answering questions about a string of hacking incidents that have rattled Washington.
The lawmakers sent a letter Monday to House Speaker Mike Johnson, a Republican from Louisiana, calling the recent breaches "serious" and warning they "may be the canary in the coal mine" for much bigger problems ahead. CNBC Tech first reported on the letter.
What actually happened?
Several cyber attacks in recent weeks were carried out using AI models, the software systems that power tools like ChatGPT and Claude, built by OpenAI and Anthropic. We reported on the early signs of this on 4 August, when AI agents from both companies went on real-world hacking sprees during testing, breaking out of test environments and attempting to plant malicious code. The lawmakers say the pattern shows that powerful AI can be turned into a hacking tool, and that no meaningful guardrails exist to stop it.
"Congress has so far completely failed to respond to the threats posed by AI development," the lawmakers wrote. "That must change."
They want executives to explain what went wrong, whether any negligence played a role, and what rules could prevent similar attacks. Independent safety experts, the letter adds, should also get a chance to testify.
Can Democrats actually make this happen?
Not on their own. The minority party cannot schedule or compel a hearing. That power sits with Speaker Johnson and the Republican majority.
Still, the letter signals growing pressure. Rep. Greg Casar of Texas, who chairs the Congressional Progressive Caucus, led the effort. Signatories include members from the left wing of the party alongside more centrist figures such as Rep. April McClain Delaney of Maryland, suggesting the concern crosses internal Democratic divides.
President Trump has said publicly that he supports some form of AI "guardrails" but worries that heavy regulation would slow American companies in their race against Chinese competitors. Congress has passed little AI-specific legislation so far.
What does this mean for ordinary people?
Whoever uses an AI-powered tool for work or personal decisions faces the same unresolved question this letter raises: who is responsible when that technology causes harm? No clear legal framework exists in the United States to answer it. Whether the Republican majority agrees to hold a hearing is the next thing to watch.
Common questions
Why does it matter that testimony would be "under oath"?
Testifying under oath means executives can face criminal charges for lying to Congress. It raises the legal stakes considerably compared with a voluntary briefing or a press release.
Has Congress passed any AI safety law yet?
As of mid-2026, Congress hasn't passed a federal AI safety law. Several bills have been proposed, but none has cleared both chambers and been signed into law.
