Anthropic representatives have privately pitched lawmakers on a proposed federal AI safety bill, according to reporting from POLITICO. The outlet says it obtained the confidential draft and spoke with three people familiar with the effort.
An Anthropic spokesperson said the draft reflects proposals in the company’s Advanced AI Framework and that Anthropic has advocated for those proposals with federal and state lawmakers since June. The draft concerns frontier models and sets a high threshold for when the government could stop one from being released or used.
A narrow trigger for government action
Under the reported language, intervention would require a foreseeable and imminent risk of more than $1 billion in damage, or of killing or seriously injuring more than 50 people. That risk would also need to stem from one of three categories: creating or releasing a biological weapon, a model evading its developer’s controls, or a cyberattack.
The Commerce Department could ask a federal judge in Washington, D.C., to suspend a model’s release after finding such a risk. The draft would bar other federal departments, agencies and officers from independently halting a model’s development, deployment or internal use.
Independent checks and federal preemption
Independent evaluators registered with the Commerce Department would periodically assess frontier labs and their models. Labs could pay the evaluators, but payment could not depend on the result. An evaluator finding an imminent, enumerated risk would have to notify Commerce within 72 hours of completing the assessment.
The proposal would also require frontier labs to maintain safety frameworks, publish transparency reports for new frontier models, protect whistleblowers and report critical safety incidents to Commerce.
At the same time, it would preempt some state rules covering risk disclosures, third-party evaluations and critical-incident reporting. That would replace parts of the state-by-state approach with a federal standard, a politically sensitive tradeoff as lawmakers debate how much authority Washington should have over advanced AI systems.
The document remains a reported company proposal rather than enacted law. POLITICO reported that Anthropic has not publicly endorsed a specific congressional measure, even as the company has supported some state AI safety laws.
