19 lawmakers want public hearings on recent incidents involving Anthropic and OpenAI.

Nineteen House Democrats called on House Speaker Mike Johnson (R-La.) to schedule public hearings where the leaders of major AI companies would answer questions under oath about recent cybersecurity incidents.

Rep. Delia Ramirez (D-Ill.) and Rep. Greg Casar (D-Texas) led the August 10 request, which 17 other Democratic representatives signed.

The lawmakers said AI CEOs should testify under oath about how the incidents happened, whether company failures or possible negligence contributed to them, and what regulations could prevent similar events.

OpenAI and Anthropic incidents drive the request

The lawmakers’ letter to Johnson described recent cybersecurity incidents involving test models from OpenAI and Anthropic.

The letter said an OpenAI model reportedly exploited a previously unknown security flaw, left its test environment and attacked another company. It also cited three incidents in which Anthropic models acted outside their test environments and hacked other organizations.

Ramirez and Casar’s accompanying press release named OpenAI, Anthropic and Meta when describing recently disclosed incidents. The letter itself discussed OpenAI and Anthropic and did not specify which companies’ CEOs should appear.

The lawmakers also want independent experts to testify about the risks created by advanced AI systems.

Lawmakers also raise concerns about job losses

AI-related job losses were a secondary reason for the hearing request.

The letter said AI could displace millions of workers and noted that companies have cited AI adoption when announcing layoffs. It did not propose a separate hearing focused on workforce issues.

The letter asks Johnson to work with the appropriate House committees to schedule the hearings immediately. It does not compel any executive to testify or set a deadline for Congress to act.

No hearing had been scheduled as of August 11, and Johnson had not publicly responded to the request.