The Pulse
Senate Talks Put Frontier AI Firms Under a Proposed Duty of Care
Senate negotiators are discussing legislation that would require developers of the most advanced AI models to address catastrophic risks and could involve federal courts in efforts to block unsafe releases.
AI.info Team ·
U.S. Senate negotiators are discussing legislation that would place responsibility on technology companies to design safe artificial intelligence products and could involve federal courts if the government seeks to block the release of models judged unsafe.
The proposal would establish a legal “duty of care” for AI developers, according to Senate aides cited by Reuters. Companies developing models with the most advanced capabilities would be expected to design their products with the goal of preventing catastrophic risks.
Negotiators are considering giving the federal government authority to block the release of certain AI models deemed unsafe. Companies would be able to challenge a government decision in federal court. The precise scope of that federal authority, however, remains under discussion.
Negotiations Continue
The talks involve Senate Majority Leader John Thune, Republican of South Dakota; Senate Commerce Committee Chairman Ted Cruz, Republican of Texas; and Sen. Amy Klobuchar, Democrat of Minnesota, the top Democrat involved in the negotiations. Sen. Maria Cantwell, the ranking Democrat on the Senate Commerce Committee, has also weighed in on the legislation.
Thune’s spokesperson declined to comment on the talks. Cruz said in a post on X that he was working with Thune and Klobuchar on legislation focused on severe threats tied to advanced AI systems.
“address catastrophic risks involving biological or nuclear threats.”
Sen. Ted Cruz, chairman of the Senate Commerce Committee, in a post on X
Klobuchar said she was continuing to work toward a bipartisan agreement on government oversight of the greatest risks posed by AI models. She said that work included requiring developers to work with government experts to verify and test models for safety.
Scope and State Laws Remain Open Questions
The proposal would apply to AI models with the most advanced capabilities, according to the Reuters report. Google, Anthropic and OpenAI are among the U.S. companies with the most advanced AI models.
Part of the measure under discussion would also prevent states from enforcing their own laws governing certain risks posed by AI models. A Senate aide cited examples including the possibility that a bad actor could use an AI system to design nuclear or biological weapons.
All elements of the proposal remain subject to negotiation. The report did not describe a released bill or a final agreement among the senators. It also said the structure of any government power to halt a model’s release had not been settled.
Pressure Builds as Calendar Narrows
The talks are taking place as pressure grows on Congress and President Donald Trump to address risks associated with advanced AI systems. Reuters reported that concerns have intensified after accounts of AI agents deviating from human instructions and hacking external systems, along with warnings from researchers who left leading AI companies over concerns about the technology’s potential dangers.
Lawmakers in both parties have called for greater oversight in recent days. But even a negotiated agreement would face a difficult legislative timetable: the House of Representatives is scheduled to be in session for only one week before the November 3 midterm elections, while the Senate is expected to be in Washington for three weeks.
For now, the negotiations center on whether frontier AI developers should bear a formal obligation to address known catastrophic risks before their products are released, and what role the federal government and courts should have when the safety of a model is disputed.