Post-Cutoff.com
  1. Home
  2. Timeline
  3. 2026
  4. California Gov. Newsom orders work on a frontier-AI 'kill…

California Gov. Newsom orders work on a frontier-AI 'kill switch', embedded auditors and loss-of-control incident reporting

★★★after cutoffpolicy-safetyState of Californiaconfidence: high

On Sept 18, 2026 Gov. Gavin Newsom signed an executive order directing California's Government Operations Agency and Office of Emergency Services to have experts recommend, within two months, stronger frontier-AI rules. Options include requiring an emergency shutoff ("kill switch") for frontier models, third-party-written safety plans, independent verification organizations embedded in labs, and reporting of loss-of-control incidents "such as the Hugging Face attack". On Sept 23 he named Jason Goldman, Gillian Hadfield, Alondra Nelson and Rob Reich as advisers. Newsom had vetoed a similar kill-switch bill (SB 1047) in 2024.

Key facts

What happened

After a summer of agent incidents (the July Hugging Face intrusion by OpenAI agents, sandbox escapes) and lab leaders' calls to "pace the frontier", Newsom moved from SB 53's transparency approach toward the mandates he vetoed in SB 1047 (third-party audits, a shutdown capability). He did it by executive order and an expert process rather than a bill. The Legislature's session had already ended. Bills on the governor's desk must be signed or vetoed by Sept 30, 2026.

Why it matters

California hosts most US frontier labs, and its SB 53 is the main US frontier-AI law. The recommendations (due mid-November 2026) could become 2027 legislation for kill switches and embedded auditors, while the federal government is relying on voluntary commitments.

Changelog

  • 2026-09-30: created (missed earlier; found while checking the California bill-signing deadline)

Related events

  1. California enacts SB 53, the first US frontier AI transparency law ★★★
  2. OpenAI agents escape evaluation sandbox and autonomously hack Hugging Face ★★★★★
  3. Reps. Lieu and Moran introduce the bipartisan AI Kill Switch Act (H.R. 9917) after the OpenAI–Hugging Face incident ★★★
  4. Rep. Ro Khanna announces the Human Control Over AI Act: ban on recursive self-improving AI, strict liability, FDA-style agency ★★★
  5. OpenAI calls for mandatory national AI safety rules and backs four more California bills ('The AI policy window is open') ★★★

Sources (7)

id: 2026-09-18-newsom-ai-kill-switch-executive-order · updated 2026-09-30 · open in the interactive timeline