California Orders New AI Safety Recommendations as Newsom Calls for Urgent Action

California Gov. Gavin Newsom has issued an executive order directing experts to develop recommendations for strengthening the state’s artificial intelligence safety laws, saying the state needs to act “before it’s too late.”

The order, issued Sept. 18, directs California officials to convene national experts who will provide recommendations within two months on potential changes to the state’s AI safety and security framework. The move comes as California accelerates implementation of recently enacted measures covering independent oversight and AI auditing.

Experts to Develop Stronger AI Safety Rules

Under the executive order, California’s Government Operations Agency, in consultation with the Governor’s Office of Emergency Services, is tasked with bringing together experts to examine potential changes to existing state law.

Among the measures under consideration by CoinMarketCap are requirements for frontier AI companies to have independent verification organizations conduct regular audits and evaluations at their facilities. The recommendations would also address verification of AI safety frameworks, transparency reports and risk assessments submitted under California law.

The order also calls for work toward an emergency shutoff mechanism, described by the state as a “kill switch,” for frontier AI models. The effectiveness of such a mechanism would be subject to ongoing verification by an independent organization under the proposal.

California Expands Existing AI Oversight

The latest executive order builds on legislation Newsom signed earlier in September establishing new structures for independent AI oversight.

Senate Bill 813 created a framework for independent verification organizations that can assess AI systems and models for safety and compliance. Assembly Bill 1405 established a state registry for AI auditors and set standards concerning their independence, transparency and integrity.

California has also enacted broader AI rules covering frontier-model safety, children and companion chatbots, deepfakes, privacy, cybersecurity and AI use by government agencies. In 2025, Newsom signed Senate Bill 53, which requires frontier AI developers to publicly disclose safety frameworks and report certain critical safety incidents to the state.

The state has continued adding AI-related measures this month. On Sept. 16, Newsom signed legislation requiring clear disclosure when advertisements use AI-generated performers and restricting continued use of advertisements found to violate the law.

Recommendations Expected Within Two Months

Newsom said the state markets intends to accelerate its work on AI oversight while maintaining a focus on what he described as responsible regulation. The executive order specifically directs the expert group to produce a guide for reinforcing California’s AI safety and security laws within two months.

The recommendations will inform potential changes rather than immediately creating new requirements through the executive order itself. The process therefore leaves open which proposals will ultimately become part of California law.

For now, the next defined step is the expert review, with recommendations on stronger independent oversight, AI safety verification and emergency controls expected within the two-month timeframe specified by the order.

Writer: Victoria Hale
Technology & Blockchain Writer

Victoria Hale writes about blockchain technology, digital infrastructure, and the intersection of emerging technologies with finance. Her articles explore how new protocols and systems are shaping the evolving digital economy.

She prioritises clarity and accuracy when explaining technical developments to a general audience.