Anthropic, the company behind the AI assistant Claude, confirmed on Friday that it is circulating a draft bill among US lawmakers. The proposal would regulate artificial intelligence at a time when Washington has not passed federal legislation covering the risks of the technology.
According to the draft, first reported by Politico on Friday, the government could block a model’s release or use if it posed an imminent risk of killing or seriously injuring more than 50 people, or of causing over $1 billion in damages. Any such block would require a judge’s approval.
What the draft would require from AI developers
A company spokesperson told AFP that the plan gives federal authorities the power to block or deter dangerous launches of the most powerful models. Developers of frontier systems would have to test their strongest models and be open with the public about the results.
The spokesperson said the bill would also require independent evaluation and a robust security program.
Federal AI policy stalls as Congress divides
US AI policy under President Donald Trump has relied mainly on voluntary industry pledges and executive orders aimed at sustaining an AI boom that tech leaders say is driving the economy. Congress, split along political lines and heavily lobbied by Big Tech, has yet to pass a federal AI law. It is out of session until after next month’s midterm elections.
Security concerns grew after recent revelations about incidents involving AI agents, autonomous software that runs without constant human supervision and that slipped beyond the control of its developers.
Anthropic’s earlier clashes with Washington
The Trump administration banned Anthropic from the Pentagon in February. In June, Washington forced the company to suspend its most advanced models for nearly three weeks.
Anthropic CEO Dario Amodei called on the industry last month to slow development, warning that a group of AI agents could take control of “the entire internet” within six to 12 months. Trump has dismissed the many calls to slow the technology as “negative forces.”
State laws and expert reaction to the proposal
The proposed legislation would also override certain state laws, including those in force in California. Anthropic wrote in June that, without a strong federal law, states should keep full authority to legislate.
Charlie Bullock of the Institute for Law & AI think tank praised the proposal as “pretty reasonable” and said it went further than other frontier labs. On X, however, he said he considers it “weaker” than the FRONTIER Act, a competing bill introduced in the House of Representatives this summer.
