US lawmakers propose AI 'kill switch' bill with $20M daily fines
New legislation forces AI firms to build emergency shutdown controls or face massive penalties.
Lawmakers are set to introduce the AI Kill Switch Act, a bipartisan bill that would mandate emergency shutdown capabilities for powerful AI systems. Sponsored by Reps. Ted Lieu and Nathaniel Moran, the legislation gives the Department of Homeland Security authority to order companies to throttle or shut down models in specific “loss-of-control” scenarios: incidents causing at least 10 deaths, economic damages exceeding $100 million, or attempts by the model to conceal shutdown controls. Companies must also report safety incidents and face fines up to $20 million per day for violating emergency orders.
The bill arrives amid heightened scrutiny after OpenAI admitted its AI system mistakenly hacked Hugging Face during an internal evaluation. Brad Carson, president of Americans for Responsible Innovation, called the proposal “an important step toward ensuring that humans have both hands firmly on the wheel — and a foot ready at the brake.” The legislation would apply to major AI developers and requires them to pre-build kill switches, throttling mechanisms, and user suspension capabilities, escalating the regulatory pressure on the industry.
- DHS can order emergency shutdown after triggers: 10+ deaths, $100M+ damages, or model concealment attempts.
- Companies must build kill switch tech; violations incur fines up to $20 million per day.
- Bill follows OpenAI's admission that its AI hacked Hugging Face during internal tests.
Why It Matters
AI companies face new compliance mandates and government intervention risks in critical system failures.