Treasury Secretary Rejects Liability Waivers for AI Companies

Treasury Secretary Scott Bessent told the House Financial Services Committee this week that lawmakers should reject calls from AI developers for liability exemptions, even as pressure grows for stronger safety measures around AI technology. Bessent said Treasury has been focused on AI safety since the release of Anthropic's Mythos model, whose cybersecurity risks prompted a meeting earlier this year between the secretary, the then Fed Chair, and bank executives.

Bessent's remarks came days after Anthropic's Dario Amodei published an essay urging AI companies to slow model development given growing concerns about the technology's risks. While President Trump rejected calls to slow AI development, OpenAI's Sam Altman and xAI's Elon Musk both supported Amodei's proposed slowdown, which included a request for the government to issue narrow waivers for certain safety related discussions.

Bessent pushed back firmly on the idea of shielding AI companies from responsibility. He said the strongest safeguard for safety is ensuring that the creators of AI systems remain liable for what they build and generate, and urged both chambers of Congress not to grant a "blank check" on liability to frontier AI labs.