- jurisdiction
- US, federal
- level
- federal
- domain
- safety
- status
- proposed
- scope
- Federal bill that would require any entity that develops or operates an advanced artificial intelligence system in the United States to include a human-controlled shutdown mechanism in that system.
- obligations
- Each covered entity must ensure the advanced AI system it develops or operates includes a technical capability for a human operator to shut down the system; the bill directs the Secretary of Homeland Security, in consultation with the heads of relevant federal agencies, to promulgate implementing regulations within 90 days of enactment.
- effectiveDate
- TBD
- appliesTo
- Any entity that develops or operates an advanced artificial intelligence system in the United States
- notes
- Introduced September 24, 2026 by Rep. Thomas H. Kean Jr. (R-NJ); referred to the House Committee on Science, Space, and Technology. Not yet passed by either chamber. Official title: to require entities to include human-controlled shutdown mechanisms in all artificial intelligence systems. Distinct from all currently tracked entities: no existing federal law mandates a human-operable shutdown capability across advanced AI systems generally; California's SB 813/AB 1405 and Executive Order N-9-26 address a kill-switch concept only for frontier models under California state law, and EO 14409's frontier-model information-sharing framework is voluntary and does not require a shutdown mechanism. Quote independently re-verified against the bill's own HTML text on GovInfo.
