California Operationalizes the Pacing Debate: Expert Panel Named for Newsom's AI Kill-Switch Order
Update (Tuesday, September 23, 2026). Governor Newsom announced the four-member expert panel that will deliver the recommendations required by his September 18 executive order (N-9-26) on independent AI oversight and a potential frontier-model "kill switch." The panel: Jason Goldman (first White House Chief Digital Officer), Gillian Hadfield (Johns Hopkins, AI alignment and regulatory design), Alondra Nelson (former acting director of the White House OSTP), and Rob Reich (Stanford, former senior advisor to the US AI Safety Institute).
The executive order directs the Government Operations Agency to accelerate implementation of SB 813 (California's first-in-nation framework certifying independent verification organizations, signed Sept 9) and AB 1405 (a state registry for AI auditors), and to develop recommendations to:
- "Require frontier AI companies to embed a designated independent verification organization onsite in their labs to conduct regular audits and evaluations."
- "Advance the creation of a 'kill switch' for frontier models, with the efficacy of the switch verified on an ongoing basis by an independent verification organization."
- "Update definitions of critical safety incidents to include loss-of-control incidents such as the Hugging Face attack" — a direct legislative response to OpenAI's rogue-agent episode (OpenAI's Rogue Agents Probed Hugging Face as Early as May — Two Accounts Hijacked During Reconnaissance).
The political framing is a direct challenge to the White House: the release opens "While Donald Trump remains unplugged, Governor Gavin Newsom continues to advance strong AI policies," notes that "the federal government has done nothing," and mocks Trump's rename of AI to "super intelligence" with a screenshot of a White House poll on the term. It demands Congress adopt California's framework "as a floor, not a ceiling."
Panel member Jason Goldman tied the order to the core verification problem: "Governor Newsom's executive order addresses the hardest questions in AI safety governance, including who verifies a frontier lab's safety claims and whether a model can be reliably shut down."
What it means. While federal policy moves in the opposite direction (AI Czar Named: DNI Jay Clayton to Chair the "Super Intelligence Force" — a 120-Day Report, Not a Regulator — an "AI Force" and czar with no regulatory teeth, and formal rejection of international governance, Labs Take the Pacing Case to the UN Security Council — and the US Rejects Global AI Governance On the Spot), California is building the actual machinery the pacing debate's proponents asked for: embedded verifiers, verified kill switches, and mandatory loss-of-control incident reporting. Nearly every frontier lab is California-based, so SB 813/AB 1405 plus this EO would put state-verified oversight inside Anthropic, OpenAI, and Google's home labs — the same embedded-evaluator concept Anthropic piloted voluntarily with Accenture (Anthropic Embeds Its First Independent Evaluators: Accenture's Faculty, $1B Each, Employee-Level Access), but compulsory. Expect the labs and Washington to fight the extraterritorial reach (the California Globe's objection: what happens when a model is "copied, fine-tuned, serving California from six outside regions").