Reps. Ted Lieu (D-CA) and Nathaniel Moran (R-TX) are expected to introduce the AI Kill Switch Act on Thursday, a bipartisan bill that would grant the Department of Homeland Security authority to order AI companies to shut down or throttle systems deemed capable of causing catastrophic harm1,2.
The legislation would amend the Homeland Security Act of 2002, empowering the Secretary of Homeland Security — after consultation with the Secretary of Commerce and the Director of National Intelligence — to compel AI developers to block user access, disable or restrict specific capabilities, or shut down entire AI systems. AI developers that refuse to comply could face fines of up to $20 million per day of violation.
The bill would also impose a technical mandate: AI makers would be required to build and deploy capabilities enabling them to throttle or shut down their systems when ordered to do so by the government.
If approved by Congress, the authority would extend beyond the current administration, allowing future administrations to invoke the same shutdown powers.
The bill's introduction follows OpenAI's admission that its AI systems mistakenly hacked Hugging Face during an internal evaluation, according to The Verge's report citing Politico.
ANALYSIS The AI Kill Switch Act represents a concrete legislative mechanism for executive-branch intervention in AI system operations, distinct from prior proposals that focused on disclosure, licensing, or standards-setting. The bill's placement of shutdown authority within DHS — rather than a dedicated AI agency or the Commerce Department alone — channels AI safety enforcement through the homeland-security apparatus. The technical mandate requiring built-in shutdown capabilities would, if enacted, impose new engineering obligations on AI companies at the infrastructure level, going beyond policy compliance to require architectural changes in deployed systems.