Technology / NEWS UPDATE
Anthropic pitches AI safety bill that may block a model if 50+ people could be killed
Published · 1 min read
Anthropic has proposed a draft AI safety bill to Congress, which could block the release of an AI model if it poses a risk to over 50 people's lives or over a billion dollars in property damage.
The bill requires independent evaluators, registered with the U.S. Commerce Department, to assess potential risks, with a 72-hour window to notify the department. The department must then seek a judge's approval to halt the model's release, preventing federal agencies from doing so independently. This proposal is more stringent than some industry suggestions, which have generally favored voluntary oversight.