Anthropic Proposes AI Safety Bill That Limits Agency Power
Anthropic's AI safety bill sets a 50‑death threshold for model bans but limits agency action without judicial approval.
Tech
Anthropic has introduced a draft policy that would allow a model to be blocked only if it could potentially cause the deaths of more than 50 people. The bill frames the measure as a safeguard against runaway artificial intelligence.
Critics note that the proposal also restricts federal agencies from halting a model's release without first obtaining a judge’s approval, effectively placing judicial oversight in front of agency action.
The dual nature of the draft—offering a high‑threshold safety trigger while curbing agency authority—has sparked debate over how AI risks will be managed under U.S. law.
Why It Matters
The proposal could shape how U.S. regulators intervene in high‑risk AI deployments, influencing global discussions on AI oversight.
More in Tech
Tech
Genetic Genealogy Boosts Cold Case Solves Amid Privacy Concerns
Tech
Judge Dismisses Sam Altman's Sister's Sexual Abuse Lawsuit
Tech
Oura Rings Add Cardio Effort and Automatic VO2 Max Tracking
Tech
TikTok Blocks Searches for Popular Peptides Including Semaglutide
Tech
AMD Rolls Out FSR 4.1 Upscaling to Xbox Series Consoles and Handhelds
Tech