Anthropic’s proposed AI safety bill only kicks in if 50 deaths or $1B in damages are on the line — proposal would effectively make the Commerce Department the only agency capable of stopping an AI model
Summary
The article reports on Anthropic's push for an AI safety bill that would allow government intervention to block AI releases only if risks exceed a $1 billion damage threshold or pose harm to 50+ people. It describes an independent evaluator under the Commerce Department and a 72-hour reporting requirement, with judicial review to halt a model release. The piece situates the proposal within broader AI safety debates and related incidents and policy discussions.