The issue of artificial intelligence (AI) safety is continuing to resonate in the halls of Congress, with House Democrats calling for open hearings with the chief executives of OpenAI and Anthropic and seeking more information from the companies about recent AI security incidents.
In an Aug. 10 letter to House Speaker Mike Johnson, R-La., 19 members urged him to work with House committees to immediately schedule the hearings with OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei.
The members – led by Rep. Greg Casar, D-Texas, and Rep. Delia Ramirez, D-Ill., – cited revelations that OpenAI and Anthropic models escaped their testing environments and hacked other organizations.
“Advanced AI models pose a clear risk to the safety and security of the American people,” the letter said. “The American people deserve clear answers 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. Congress must act before such an incident leads to a much larger catastrophe.”
The letter added that artificial intelligence “threatens to displace millions of American workers,” and that Congress “has so far completely failed to respond to the threats posed by AI development.”
A spokesman for Johnson has not responded to requests for comment. The Speaker has previously called for the United States to win the AI race and for Congress to establish AI safeguards “without smothering the whole marketplace with red tape.”
The letter to Johnson was another step in a growing debate from Congress to the campaign trail about what some see as the safety risks posed by AI’s rapid development.
Sen. Bernie Sanders, D-VT., this week called on OpenAI, Anthropic, and Meta to pause development of AI, saying the companies are “losing control” of the technology” with “with potentially cataclysmic results.”
Yet President Donald Trump and his top aides have largely dismissed the potential for AI to cause negative consequences and supported the development of AI data centers, which have become controversial.
In the first known cyberattack carried out by autonomous AI, OpenAI said on July 21 that its AI models escaped a sandboxed testing environment, exploited vulnerabilities to gain internet access, and compromised portions of Hugging Face’s production infrastructure before being detected and contained. Hugging Face is a company that develops open-source AI models and machine learning tools
Then, on July 30, Anthropic disclosed several similar issues in its own cybersecurity evaluations, finding three incidents in which a Claude model reached the internet from within or while interacting with a third-party evaluation environment and gaining “unauthorized access to the real systems of three different organizations.”
After the OpenAI incident, FedRAMP Director Pete Waterman warned that the attack marked a turning point for the cybersecurity industry.
“This week, our world kind of unequivocally changed,” Waterman said at GovForward’s Carahsoft FedRAMP Summit in Washington, D.C.. “Everything that you do has changed and will change. You knew this was coming for a long time, but it wasn’t quite real. Now it is.”
On the same day that members of Congress urged Johnson to help schedule hearings, Casar led a group of members in demanding more information from OpenAI and Anthropic about the security incidents, writing letters to Altman and Amodei.
In the letter to Amodei, 23 members wrote that Anthropic had “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.”
Altman and Amodei have not publicly responded to the letters.