TL;DR
California Governor Newsom has signed two AI safety bills and appointed a four-person expert panel to design independent oversight mechanisms, including a mandatory "kill switch" for frontier AI models, positioning the state as the de facto national regulator while federal action stalls.
What happened
- Governor Newsom signed SB 813 (McNerney), creating a framework for independent verification organizations to assess AI systems and models for safety and risk.
- AB 1405 (Bauer-Kahan) also signed: establishes a state registry for AI auditors assessing compliance with California law.
- A four-person advisory panel was named to guide implementation of Newsom's AI executive order, issued days before the announcement.
- Two core proposals are under active consideration: embedding independent third parties inside frontier AI companies to verify safety claims, and requiring companies to build an emergency shutoff, or "kill switch," for frontier models.
- Advisors include Rob Reich (Stanford, former US AI Safety Institute senior advisor), Gillian Hadfield (Johns Hopkins, Vector Institute), Alondra Nelson (Institute for Advanced Study, former acting White House OSTP director), and Jason Goldman (first White House Chief Digital Officer).
Why it matters
- California is filling the federal vacuum: with the Trump administration disengaged on AI safety, the world's fifth-largest economy is setting standards that frontier labs, most headquartered in California, will have to meet.
- The "kill switch" mandate would be a first: no jurisdiction has yet required frontier model developers to build and maintain a verified emergency shutoff, making this a potential global precedent.
- Independent embedment inside labs is the most aggressive proposal on the table: third-party verifiers inside companies like OpenAI or Anthropic would fundamentally alter how safety evaluations are conducted and disclosed.
- SB 813 and AB 1405 together create infrastructure: a registry of certified auditors plus a verification-organization framework gives California the institutional machinery to enforce whatever the advisory panel recommends.
- The panel's academic and policy pedigree (two former White House officials, two leading AI governance scholars) signals these recommendations will carry weight in Washington and Brussels, not just Sacramento.
What to watch next
- Advisory panel timeline: watch for the panel's formal recommendations, which will determine how aggressive the final kill-switch and embedment rules actually are.
- Industry response: whether frontier labs (OpenAI, Anthropic, Google DeepMind) comply quietly, lobby hard, or litigate will signal how enforceable California's framework becomes in practice.
- Federal preemption risk: any move by Congress or the Trump administration to preempt state AI regulation would directly undercut this entire framework, making federal legislative activity the key variable to track.
Originally published on Present of AI, a daily source-linked AI news timeline. Read the full timeline or browse the open dataset.