California's AI auditor registry (AB 1405) and the kill-switch executive order

What the three documents are

This guide draws on three pages from the Office of Governor Gavin Newsom, dated September 9, September 18 and September 23, 2026. They are press releases. They are not the text of the bill or of the order. None of the three pages uses the designation "N-9-26," so this guide calls it the September 18 executive order. Every description below is the Governor's office's own account and is attributed that way.

What AB 1405 establishes

The September 9 page says the Governor signed two bills that day, Senate Bill 813 and Assembly Bill 1405.

The page describes AB 1405, authored by Assemblymember Rebecca Bauer-Kahan, as creating a state registry for AI auditors and establishing standards for their independence, transparency, and integrity.

Bauer-Kahan is quoted on the page explaining the purpose: "We cannot expect industry to simply grade its own homework; third-party auditors are essential to ensuring AI is safe for our communities and critical infrastructure."

The pages do not say what the independence, transparency and integrity standards contain. They also do not say who must use a registered auditor, how an auditor gets onto the registry, or what penalties apply.

How SB 813 sits next to it

The September 9 page says Senator Jerry McNerney authored SB 813, which establishes a first-in-the-nation framework for independent verification organizations that can assess AI systems and models for compliance with state law.

The September 18 page adds that the framework certifies those organizations. It says California is the first state to establish a framework for certifying independent verification organizations with sufficient expertise and demonstrated independence from AI companies to objectively assess AI systems and models for safety and risk.

In the pages' description, then, AB 1405 covers a registry of auditors and SB 813 covers certified verification organizations.

For context, the September 9 page says the 2025 law SB 53 requires frontier AI developers to publicly disclose their safety frameworks, report certain critical safety incidents to the state, and protect whistleblowers.

What the September 18 executive order directs

The September 18 page says the order directs the Government Operations Agency to accelerate the implementation timelines for SB 813 and AB 1405.

The page says the agency is also to work with the Governor's Office of Emergency Services to convene national experts who will develop recommendations for potential changes to state law.

The page lists four areas for those recommendations:

• The first is a requirement that frontier AI companies embed a designated independent verification organization onsite in their labs to conduct regular audits and evaluations.
• The second is a requirement that the safety frameworks, transparency reports and risk assessments frontier companies must file under state law be verified against standards an independent verification organization deems adequate.
• The third is to advance a "kill switch" for frontier models, with an independent verification organization checking the switch's efficacy on an ongoing basis.
• The fourth is to update the definitions of critical safety incidents to include loss-of-control incidents such as the Hugging Face attack.

The page says the order convenes a group of experts to provide, within two months, a guide for California to reinforce and strengthen its AI safety and security laws.

The page also says that no federal law requires AI companies to report dangerous incidents when they happen.

Who is advising

The September 23 page says the Governor announced the group of experts who will advise on the recommendations required by the order. It names four people, described as follows:

• Jason Goldman is a board member of the Center for Shared AI Prosperity, an early product leader at Google and Twitter, and the first White House Chief Digital Officer.
• Gillian Hadfield is a professor at Johns Hopkins University who focuses on AI alignment and the design of legal, regulatory and technical systems for AI.
• Alondra Nelson is a professor at the Institute for Advanced Study, director of the Science, Technology, and Social Values Lab, and a former acting director of the White House Office of Science & Technology Policy.
• Rob Reich is a professor at Stanford University who focuses on the governance of frontier science and technology, and a former senior advisor to the United States AI Safety Institute.

The page presents these four as advisers on the recommendations. It does not say they have any power to decide.

What the pages leave open for security leaders

The three pages contain no instructions for security leaders and no compliance dates for companies. They do point to three developments: the expert group's guide, the accelerated timelines for SB 813 and AB 1405, and the updated definition of critical safety incidents. The pages leave these questions unanswered:

• Dates. The new implementation timelines for SB 813 and AB 1405 are not stated. The expert guide is described only as due within two months of the order.
• Scope. All four recommendation areas name frontier AI companies. The pages do not say whether organizations that use or buy frontier models would face any requirement.
• Incident definitions. The pages do not define "loss-of-control incident" and do not describe the Hugging Face attack.
• The kill switch. The pages say it is to be advanced and verified. They do not say what it is technically or who would operate it.
• The registry. The pages do not give the content of the auditor standards or the process for getting on the registry.

Until the order's text, the expert guide or the implementing rules are published, these pages support no statement about specific obligations.