On 16 October, Reuters reported that Anthropic representatives had pitched Capitol Hill lawmakers on a draft frontier AI safety bill, relaying Politico’s account based on three people with knowledge of the effort. The draft would allow the government to stop a model from being released or used only when the threat is both foreseeable and imminent: damages exceeding $1 billion, or death or serious injury affecting more than 50 people. That risk would have to arise from the creation or release of a biological weapon, a model evading its developer’s controls, or a cyberattack.
Independent evaluators would register with the Commerce Department and periodically assess a laboratory and its frontier models. Labs could pay them, provided payment did not depend on the outcome. Evaluators finding an imminent risk in the specified categories would have to notify Commerce within 72 hours of completing their assessment.
Commerce could ask a judge to halt a model after determining risk. The draft would prohibit other US government departments, agencies and officers from suspending development, deployment or internal use. A federal judge for the District Court for the District of Columbia could order a model’s suspension.
