Gov. Gavin Newsom’s Sept. 18 executive order gives California agencies 60 days to recommend embedded on-site auditors at frontier AI labs, third-party verification of safety frameworks, and an emergency shutoff for frontier models. Recommendations are due Nov. 16, and they’ll land on top of an AI regulatory stack that’s already the strictest in the country.

The order directs the Government Operations Agency, in consultation with the Governor’s Office of Emergency Services, to accelerate implementation of SB 813 (McNerney) and AB 1405 (Bauer-Kahan). It also expands the definition of critical safety incidents to include loss-of-control events like the Hugging Face attack, and revives structural elements Newsom rejected two years ago when he vetoed SB 1047 in 2024: third-party auditing, a kill switch, clearer legal liability.

The politics are legible. SB 53, enacted in 2025 as the Transparency in Frontier Artificial Intelligence Act, already defines catastrophic risk to include incidents involving 50 or more deaths, chemical or biological weapons, or more than $1 billion of theft or damage. Per CalMatters, the November recommendations could seed a special legislative session, a step Newsom floated this week in an interview with Politico. The governor who killed SB 1047 is now sequencing its return through executive channels.

For small businesses running AI-generated ads, chatbots, or pricing agents in California, the compressed timeline hits the vendor layer, not the end-user. Before mid-November, owner-operators should confirm the AI tools they use for customer acquisition are sourced from vendors registered under AB 1405, that any AI-generated ads carry the disclosures SB 1050 now requires, and that customer-facing chatbots meet Adam’s Law risk-assessment mandates.

The auditor registry isn’t theoretical anymore. It’s a Nov. 16 deliverable.

Sources

Sources