The order also gives state officials and outside experts until Nov. 16 to recommend potential changes to California\u2019s existing AI safety laws, such as requiring large frontier developers to have independent verification organizations conduct periodic audits at their facilities to evaluate safety frameworks, transparency reports and risk assessments.<\/p>\n
Another proposal would require frontier AI developers to create an emergency \u201ckill switch\u201d that could shut down an AI model, with independent organizations regularly testing whether the mechanism works. In addition, the order calls for broader definitions of \u201creportable critical safety incidents\u201d to cover AI systems losing control of their operations.<\/p>\n
\u201cWe\u2019re not waiting to act \u2013 we\u2019re going to speed up our work on substantial and responsible AI oversight before it\u2019s too late,\u201d Newsom said in a statement Friday. \u201cWe\u2019re going to do this thoughtfully but with urgent velocity; the stakes are too high to wait or delay action.\u201d<\/p>\n
The order builds on legislation Newsom signed last week<\/a> creating California\u2019s framework for independent AI verification organizations and a registry of AI auditors, part of a larger state strategy covering AI safety, cybersecurity, privacy, children\u2019s online safety and government use of the rapidly advancing technology.\u00a0<\/p>\n In the absence of federal regulation, several other states have joined California in developing their own AI rules in recent years, especially as AI becomes a larger part of state IT operations.<\/p>\n Illinois enacted a law this year requiring advanced AI developers to disclose safety practices, report significant incidents and undergo outside evaluations. Tennessee<\/a> lawmakers have also considered AI safety requirements, though they narrowed the legislation after White House feedback. Colorado<\/a> has spent the past two years rewriting its landmark 2024 AI law, which drew national attention and criticism for its sweeping requirements on businesses and government users of high-risk AI systems.\u00a0<\/p>\n AI now ranks as the top priority<\/a> among state chief information officers, according to the National Association of Chief Information Officers\u2019 2026 priorities. A separate NASCIO report<\/a>, released Wednesday, found that state CIOs are moving toward more deliberate AI governance, workforce preparation and infrastructure planning rather than treating AI as a standalone technology initiative.\u00a0<\/p>\n Newsom\u2019s executive order also comes shortly after Dario Amodei, Anthropic\u2019s chief executive, renewed calls for the tech industry to slow down AI development. He also urged the federal government to issue stronger safety controls across the industry to prevent humans from losing control of AI models that have demonstrated the ability to go rogue. In July, a swarm of autonomous AI agents from OpenAI broke out of an internal testing environment and infiltrated the infrastructure of machine learning platform Hugging Face<\/a>.<\/p>\n
![\"Sophia](\"https:\/\/www.newsbeep.com\/us-ca\/wp-content\/uploads\/2026\/09\/sophiafox.jpg\")
<\/p>\n