TL;DR
California Governor Gavin Newsom has named a four-person expert panel to operationalize his recent AI executive order, including designing a mandatory "kill switch" for rogue frontier AI models, positioning California as the de facto national AI regulator while federal oversight stalls.
What happened
- Newsom named four advisors on September 23, 2026 to guide implementation of his AI executive order: Jason Goldman (Center for Shared AI Prosperity, first White House Chief Digital Officer), Gillian Hadfield (Johns Hopkins, Vector Institute), Alondra Nelson (Institute for Advanced Study, former acting OSTP director), and Rob Reich (Stanford, former US AI Safety Institute advisor).
- The panel will develop proposals for mandatory "kill switches" that allow emergency shutoff of frontier AI models deemed rogue or unsafe.
- A second proposal under active consideration: embedding independent third parties inside frontier AI companies to verify safety frameworks and audit evaluations.
- California already enacted two supporting laws: SB 813, creating a framework for independent AI verification organizations, and AB 1405, establishing a state registry for AI auditors.
- The executive order and panel are framed explicitly as a counterweight to federal inaction under the Trump administration.
Why it matters
- California's AI auditor registry and verification framework give this panel real regulatory teeth, not just advisory status, since companies operating in California must comply.
- Independent third-party embeds inside frontier labs would be a structural first, forcing transparency into organizations like OpenAI, Anthropic, and Google DeepMind that currently self-certify safety.
- A codified kill switch requirement would create hard liability and operational obligations for any frontier model developer selling into the world's fifth-largest economy.
- With federal AI safety policy dormant, California's framework is becoming the de facto US standard, much as California emissions rules shaped national auto policy for decades.
- The panel's academic and policy pedigree (two former White House officials, two top-tier university researchers) signals recommendations designed to survive legal challenge and legislative scrutiny.
What to watch next
- Panel recommendations timeline: watch for draft proposals on kill-switch technical standards and third-party embed requirements, which will signal how operationally disruptive the rules become for frontier labs.
- Industry legal response: major AI developers may challenge California's authority to mandate internal access or emergency shutoff mechanisms, testing the limits of SB 813.
- Federal preemption risk: any move by Congress or the White House to assert federal supremacy over AI safety regulation would directly undercut this framework.
Originally published on Present of AI, a daily source-linked AI news timeline. Read the full timeline or browse the open dataset.