Anthropic privately pitches lawmakers its own AI safety bill

On 16 October, Politico reported that Anthropic has privately pitched US lawmakers a draft federal bill on the safety of the most advanced AI. Under the text, which the paper obtained, the government could block a model only for a foreseeable, imminent risk of more than $1 billion in damage or more than 50 deaths or serious injuries, from a biological weapon, a cyberattack or the model escaping its developer's control.

Independent evaluators, whom the labs could pay, would review the models, and only the Commerce Department could ask a judge to halt one. The bill would override some state rules. Anthropic says it reflects proposals it has been advocating since June.