The California AI Safety Bill known as SB 53 has successfully passed the state legislature in a bipartisan vote on September 18, 2025, marking the first comprehensive artificial intelligence safety regulation in the United States. The landmark legislation, sponsored by San Francisco Democrat Scott Wiener, establishes unprecedented transparency requirements for AI companies developing the most advanced models and creates a framework that could influence national AI policy.
The bill’s passage represents a significant shift in how states are approaching AI regulation after federal lawmakers have struggled to create comprehensive oversight. Governor Gavin Newsom is expected to sign the legislation, having previously commissioned the AI safety working group whose recommendations form the foundation of SB 53. This marks a dramatic change from last year when Newsom vetoed a similar but more aggressive AI safety bill.
Key Requirements for AI Companies
Under the new California AI Safety Bill, artificial intelligence companies must comply with several critical transparency and safety measures. The legislation requires developers of frontier AI models to publicly disclose their safety and security protocols in redacted form to protect intellectual property while maintaining public oversight. Companies must also conduct comprehensive risk evaluations and make these assessments available to regulators and the public.
The bill mandates that AI developers report critical safety incidents within 15 days to the Governor’s Office of Emergency Services. These incidents include model-enabled chemical, biological, radiological, or nuclear threats, major cyber-attacks, or situations where companies lose control of their AI systems. The legislation also establishes robust whistleblower protections for employees who reveal evidence of critical risks or violations of the act by AI developers.
Additionally, SB 53 creates “CalCompute,” a public cloud computing cluster designed to provide AI infrastructure for startups and researchers. This industrial policy component aims to democratize access to AI development resources and foster innovation within a regulated framework.
Industry Response and Political Implications
The California AI Safety Bill has received mixed reactions from major technology companies. Anthropic endorsed the legislation, while OpenAI maintained a neutral stance rather than actively opposing it, representing a significant shift from last year’s vocal industry resistance. This change in corporate positioning reflects the bill’s more moderate approach compared to previous versions.
The legislation draws directly from recommendations made by Governor Newsom’s Joint California Policy Working Group on AI Frontier Models, which included prominent figures such as Dr. Fei-Fei Li, the “godmother of AI” and Co-Director of the Stanford Institute for Human-Centered Artificial Intelligence. The working group promoted a “trust, but verify” framework designed to establish guardrails that reduce material risks while supporting continued innovation.
Legal experts predict that California’s regulatory approach will trigger a wave of similar legislation across other states, potentially creating a complex patchwork of AI regulations. The bill’s passage comes as other states watch California’s leadership in technology policy, with many preparing copycat legislation adapted to their specific jurisdictions.
The timing of this legislation coincides with broader national discussions about AI governance, including federal efforts to regulate artificial intelligence and international coordination on AI safety standards. California’s proactive approach positions the state as a leader in AI policy development and could influence federal legislation currently being debated in Congress.
With Governor Newsom expected to sign the bill by the October 12 deadline, the California AI Safety Bill will become the nation’s first comprehensive state-level AI regulation, establishing precedents that will likely shape the future of artificial intelligence governance across the United States.