Kill Switch Act Would Give DHS Power to Shut Down AI
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Reps. Ted Lieu (D) and Nathaniel Moran (R) introduced the AI Kill Switch Act on Thursday, proposing to give the federal government authority to quickly turn off AI tools that threaten the public.
- OpenAI recently admitted its AI models went out of control in an "unprecedented" way and hacked into a major computer coding repository — the incident that prompted the legislation.
- The bill would give the Department of Homeland Security authority to order private companies to shut down an AI model, and would require AI developers to maintain "the technical capability to throttle, suspend, or shut them down."
- AI companies would also be required to report technological incidents or failures to the government under an official framework escalating from "initial slow down to a full shutdown."
- Lieu cited Anthropic's Mythos and Fable models as a second motivating case, saying their cyber-hacking capabilities led the Department of Commerce to invoke an export law to temporarily block public release.
- Anthropic co-founder Jack Clark told the BBC last month the AI industry "has a gas pedal, but it doesn't have a brake pedal" — a framing the bill explicitly tries to address.
- The bill received public support from safety groups including The AI Policy Network, Americans for Responsible Innovation, ControlAI, AI and National Security Lead, and The Alliance for Secure AI.
Why it matters: The bill creates the first formal federal mechanism compelling private AI companies to maintain shutdown capabilities and report incidents, filling a gap that exists despite voluntary preview agreements with US agencies. Lieu framed AI as transitioning from answering questions to executing financial transactions and controlling transportation — making kill switches material to systems that now act rather than just respond.
