Anthropic proposes federal safety bill with court-ordered halts for high-risk AI models

The draft, which Politico reported first and an Anthropic spokesperson confirmed, sets two triggers for a block: an imminent risk of more than 50 deaths or over $1 billion in property damage. The risk must stem from a biological weapon, a model evading its developer's controls, or a cyberattack. Independent evaluators registered with the Commerce Department would have 72 hours to alert the agency, which would then ask a Washington judge to halt distribution. Federal agencies could not pause a release on their own. The thresholds mirror those in California's SB 53, a state law on catastrophic AI risk. The draft does not explain how lab-paid evaluators would stay independent, since the bill only requires that payment not depend on their findings. Anthropic says the text reflects its Advanced AI Framework. For founders, the proposal offers a possible template for release gating. If it advances, teams shipping frontier-adjacent models could face registered third-party evaluations, incident reporting and the prospect of a court-ordered pause. Evaluator rules and the 72-hour clock deserve close reading before release processes get designed around them.