California Gov. Gavin Newsom signed sweeping artificial intelligence (AI) safety legislation into law Tuesday, making it the first state to require major AI developers to meet mandatory safety standards and disclose potential risks from cutting-edge systems.
The Transparency in Frontier Artificial Intelligence Act, or Senate Bill 53, also expands whistleblower protections for employees and executives who report unsafe practices or potential harms. While other states have passed AI-related regulations, California’s law breaks new ground by specifically targeting the safety of advanced AI products.
The legislation, authored by state Democratic Sen. Scott Wiener of San Francisco, follows months of contentious debate over the role of state regulation in the rapidly evolving AI industry. Newsom previously vetoed a similar bill, SB 1047, before signing this revised version.
The law’s impact extends far beyond California’s borders. The state hosts the world’s largest concentration of AI companies, including NVIDIA Corp., Meta Platforms Inc., OpenAI, Alphabet Inc.’s Google, and Anthropic. A recent Forbes analysis found 32 of the top 50 AI firms operate in or near Silicon Valley.
“This is a blueprint for well-balanced AI policies beyond our borders, especially in the absence of a comprehensive federal AI policy framework,” Newsom, who has emerged as the most forceful Democratic voice against President Donald Trump, said in a statement.
The governor emphasized that California has “proven that we can establish regulations to protect our communities while also ensuring that the growing AI industry continues to thrive. This legislation strikes that balance.”
California’s AI safety law stands in stark contrast to the efforts of the Trump administration, which unsuccessfully tried to erase state AI laws for a decade under the One Big Beautiful Bill Act passed in July.
The Transparency in Frontier Artificial Intelligence Act has reignited tensions between advocates of state-level oversight and those favoring federal regulation. Critics warn that a patchwork of state laws could hamper U.S. competitiveness against China in the global AI race.
On Tuesday, Sens. Josh Hawley, R-Mo., and Richard Blumenthal, D-Conn., introduced federal legislation requiring AI companies to evaluate advanced systems and collect data on potential adverse incidents.
Reactions from AI companies have been cautious. An OpenAI spokesperson called the law “a critical path toward harmonization with the federal government.” He described federal oversight as “the most effective approach to AI safety.”
Jack Clark, co-founder and policy chief at Anthropic, offered qualified support, praising the law’s “meaningful transparency requirements for frontier AI companies without imposing prescriptive technical mandates.”
Wiener defended the legislation as a necessary response to emerging threats. “California is stepping up to meet potential present and future safety issues head on,” he said, adding the law brings “commonsense guardrails” based on trust, fairness and accountability.
Ryan Hauser, a research fellow with the Mercatus Center, characterized California’s legislation as a “modest effort to promote transparency in AI development” while criticizing federal lawmakers for failing to establish national standards in the face of an impending government shutdown.
Hauser argued executive action alone cannot provide the stable regulatory framework needed for the technology sector. Presidential executive orders remain vulnerable to reversal by future administrations, he noted, creating persistent uncertainty for companies navigating the AI landscape.
An absence of federal legislation has forced technology firms to comply with a patchwork of state-level requirements from California, New York, Texas, and Tennessee, he said. This fragmented approach places companies in “perpetual limbo” as they attempt to reconcile divergent regulatory standards across jurisdictions, according to Hauser.
