Anthropic pitched lawmakers a draft bill that would let the government block a frontier AI model only if it posed an imminent risk of over $1 billion in damages or more than 50 deaths or serious injuries.
The bill reflects Anthropic’s Advanced AI Framework, an Anthropic spokesperson told Politico, which obtained the draft and reported, citing three people familiar with the matter. “These are sensible policies we believe Congress should act on,” the spokesperson added.
Anthropic did not immediately respond to Benzinga’s request for comment.
How Anthropic’s Draft Bill Would Regulate AI Models
To qualify, the risk must come from a biological weapon, a model evading its developer’s controls or a cyberattack, according to the draft. That language tracks California’s AI laws, including SB 53, signed into law last year, the report noted.
Under this bill, independent evaluators registered with the Commerce Department would assess frontier models, the most advanced AI systems. Labs could pay them, but not based on results. Evaluators must report imminent risks within 72 hours, according to the report.
Only a federal judge for the U.S. District Court for the District of Columbia could suspend a model, on a request from the Commerce Department. Other agencies would be barred from halting development or use.
Draft Bill Could Reshape AI Oversight and State-Level Rules
The bill would also preempt some state AI laws, meaning federal rules would override them on risk disclosure, outside evaluation, and incident reporting. Democrats have resisted such proposals, the report noted.
The pitch follows export controls that shut down two of Anthropic’s advanced models earlier this year.
Other AI safety proposals are also moving in Congress. The House FRONTIER Act, led by Reps. Lori Trahan (D-Mass.) and Jay Obernolte (R-Calif.), would let the government stop deployment of risky models. In the Senate, Majority Leader John Thune and Sens. Amy Klobuchar and Ted Cruz are negotiating a bill to mitigate catastrophic AI risks.
Last month, Anthropic researcher Evan Hubinger put the odds of AI wiping out humanity within a decade above 10%, amid a broader industry safety debate.
Disclaimer: This content was partially produced with the help of AI tools and was reviewed and published by Benzinga editors.
Image via Shutterstock
Recent Comments