Anthropic
Anthropic pitches Congress a bill letting a court halt AI models
Anthropic has shown lawmakers a draft federal bill that would let one court suspend a frontier model and would override state AI rules.
Anthropic has pitched lawmakers a draft federal bill that would let a court halt a frontier AI model that poses an imminent risk of mass harm. The draft would also replace state rules on the same subject.
Anthropic Confirmed the Pitch to Capitol Hill
Anthropic representatives have quietly pitched lawmakers a draft bill to regulate frontier AI safety and security, Politico reported Oct. 16, citing three people with knowledge of the efforts. An Anthropic spokesperson confirmed the substance to Reuters and Bloomberg.
The spokesperson told Reuters the bill reflects proposals in the company’s “Advanced AI Framework,” released earlier this year. It would give the government the “ability to block or deter dangerous deployments of the most powerful models,” the spokesperson said.
A Federal Court Would Have the Final Say
The government could block a model only if it poses a foreseeable, imminent risk of more than $1 billion in damages, or of killing or seriously injuring more than 50 people, Politico reported, as relayed by Reuters. The risk must come from a biological weapon, the model evading its developer’s controls, or a cyberattack.
Commerce could ask a judge to halt a model, and a judge on the U.S. District Court for the District of Columbia could order the suspension. The draft explicitly bars any other department, agency or officer from suspending development, deployment or internal use.
Independent evaluators registered with Commerce would periodically assess each lab and its frontier models. An evaluator that finds an “imminent enumerated risk” would have 72 hours after finishing the assessment to report it to Commerce.
The Draft Would Override State Rules
The draft would override state and local rules on public risk disclosures, third-party audits and incident reporting, Business Today reported Oct. 17. Reuters reported the bill would override certain state laws, such as those in force in California. Anthropic wrote in June that “absent a strong federal law, states should retain full authority to legislate,” Reuters reported.
Preemption is disputed in Congress. Business Today said Democrats largely resist weakening state safeguards, while technology groups argue that differing state rules create excessive compliance burdens. Congress has yet to pass a federal AI law and is out of session until after next month’s midterm elections, Reuters reported. Commerce did not immediately respond to a Reuters request for comment.
An Analyst Called It Reasonable but Weaker
Charlie Bullock of the Institute for Law & AI called the proposal “pretty reasonable” and “more than other frontier labs have done,” Reuters reported. He said on X that he considers it “weaker” than the FRONTIER Act, a competing bill introduced in the House this summer.
What It Means for the Labs
We think the draft trades broad, unpredictable intervention for a narrow, court-supervised one. The bar for a federal halt is high. One court has the final say. Other agencies lose the power. For Anthropic’s competitors, it sets the terms they would be measured against if Congress takes it up.