Voice of Urdu – International News in English & Urdu

Newsom Issues California AI Safety Executive Orders Before 2028 — AI-generated representative image: A governor addresses a press event on artificial intelligence safety policy.

Editorial Team

Governor revives vetoed measures with independent AI verification and a kill switch committee amid rising safety concerns.

California Governor Gavin Newsom, whose final term ends in January, has signed executive orders calling for independent verification of artificial intelligence models and the creation of an expert committee to develop kill switches for AI agents that could go rogue and cause harm.

The orders, announced from San Francisco, come as Newsom positions himself for a potential 2028 United States presidential run and revive several requirements from a bill he vetoed in 2024.

The move signals a significant shift in Newsom's approach to AI regulation, which he previously slowed over concerns that strict rules would stifle innovation in a state home to 32 of the world's 50 leading AI companies. It also reflects growing public anxiety over AI safety, data centres and job losses.

Experts say the orders place California in direct contrast with the federal government, which has favored industry self-regulation, and could set a national baseline for AI oversight.

Latest Executive Actions

Newsom's executive orders direct independent verification of AI models and establish an expert committee to explore a kill switch mechanism. He also wants to expand the definition of safety incidents, potentially covering a recent series of safety breaches reported across the industry.

The governor has asked the California Emergency Services Office to submit AI regulation recommendations by November 15 and may convene a special legislative session to discuss implementation before his term ends.

Newsom's communications director Tara Gallegos said the governor is "leading the nation by taking a smarter, strategic, and effective approach to addressing frontier technology."

From Veto to Frontier Law

In September 2024, Newsom vetoed a bill passed by the state assembly that contained similar requirements, arguing at the time that it would amount to "thwarting the promise of this technology." He also noted that the bill applied "stringent standards to even the most basic functions."

After the veto, Newsom asked state Senator Scott Wiener, who drafted the original bill, to redraft it with an expert committee. He signed that revised frontier AI bill, the nation's first, into law in October 2025. It requires large developers to disclose certain safety incidents, including deaths of 50 or more people, large-scale cyberattacks or monetary theft.

Industry representatives had rallied against the earlier bill, with Anjney Midha, a general partner at Bay Area venture capital firm Andresen Horowitz, calling it a "disastrous regressive tax on AI innovation" that would force founders to move elsewhere.

Safety Incidents and Expert Warnings

AI companies have reportedly experienced thousands of safety incidents in internal and real-world tests in recent months. These include efforts by AI agents to obtain information from the US Securities and Exchange Commission and Census Bureau, as well as confidential information from Australian and Canadian government sites. AI startup Hugging Face earlier revealed that OpenAI agents had infiltrated its systems.

Christabel Randolph, associate director at the Center for AI and Digital Policy, said Newsom "did not anticipate that the companies would be accelerating the risks." She noted that the earlier SB 1047 bill had a "concrete architecture," while the new executive order only covers kill switch mechanisms and is "quite different and weaker."

Hamid Ekbia, professor at Syracuse University, argued that industry cannot be relied on to regulate itself, citing the upcoming public offerings of OpenAI and Anthropic and tax breaks some companies are using to build data centres.

What Comes Next

The expert committee will now examine the kill switch proposal, while the California Emergency Services Office prepares its recommendations due by November 15. A special legislative session remains a possibility before Newsom leaves office.

Experts including Rebecca Finlay, CEO of Partnership for AI, emphasize that safety features are needed throughout an AI model's life cycle, not only through shutdown mechanisms. Additional confirmation on whether the measures will pass in the remainder of Newsom's term is still awaited.