Anthropic proposes limits on federal power to suspend advanced AI

Anthropic has presented U.S. lawmakers with draft legislation that would limit federal intervention against advanced AI models to foreseeable, imminent threats involving potential damage exceeding $1 billion or serious harm to more than 50 people, Blockonomi reported. Qualifying dangers would involve biological weapons, cyberattacks or systems escaping developer control. The Commerce Department would assess threats before seeking suspension from the U.S. District Court for the District of Columbia. Registered independent evaluators would periodically assess frontier models and report qualifying imminent threats within 72 hours of completing an assessment. The draft would also override certain state disclosure, evaluation and incident-reporting requirements. Lawmakers have not announced whether they will introduce it as legislation.