GoKawiilAnthropic pushes draft AI bill setting 50-death threshold for shutdown order
Politico reports that Anthropic representatives have been privately negotiating with lawmakers over draft federal AI legislation. Sources describe a proposal in which the government could only order an AI model shut down if an independent monitor, overseen by the Commerce Department, determines it risks killing 50 or more people, causing over $1 billion in damages, or triggering a bioweapon release, major cyberattack, or loss of control — with a federal judge making the final call within 72 hours of a report.
GoKawiil's interpretation of the reporting above, not reported fact.
The layered approval process — monitor, Commerce officials, then a judge — suggests Anthropic favors a high, narrowly defined bar before any government intervention, which could slow or limit regulators' ability to act quickly against emerging AI risks. The lobbying effort comes amid tens of millions in AI industry political spending ahead of US midterms, which may raise questions about how much influence companies are exerting over the specific wording of rules meant to govern them.
- Anthropic reportedly wants a 50-death or $1 billion damage threshold before the government can force an AI model's shutdown.
- The proposed process involves a third-party monitor, Commerce Department officials, and ultimately a federal judge's approval within 72 hours.
- The lobbying is part of a broader wave of AI industry political spending, totaling at least $55.7 million, ahead of the US midterms.
Source: futurism.com — Joe Wilkins, 2026-10-19
Published there as: “Anthropic Secretly Lobbying Government to Only Shut Down an AI Model If It’s Poised to Kill More Than 50 People”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.