Looking ahead: California's new law cements the state's role at the forefront of the national debate. By enshrining transparency and reporting requirements into statute, it establishes a precedent that other states and federal lawmakers will now have to confront as they consider how far to go in regulating the technology.
California has become the first state to require major artificial intelligence companies to make their safety practices public. Governor Gavin Newsom signed the bill, known as SB 53, after months of debate between lawmakers and technology firms including OpenAI, Meta, and Anthropic. The law is already attracting attention in Washington and abroad, as Congress and foreign governments weigh how to regulate a field that has advanced faster than policymakers have been able to respond.
The legislation marks a policy shift for Newsom, who last year vetoed a more expansive AI safety bill. That earlier proposal would have imposed sweeping restrictions and faced fierce resistance from Silicon Valley. This time, the bill's scope was narrowed, and legislators worked with both the governor's office and major companies to address some of the industry's concerns.
Under SB 53, advanced AI companies must publish their safety and security protocols and notify the state of safety incidents. These include criminal uses of AI without human involvement, such as autonomous cyberattacks, or instances where a model behaves deceptively or manipulatively. The law also provides a pathway for companies and individuals to report risks, with protections in place for whistleblowers.
Additionally, the measure sets in motion plans for a state-run cloud computing project, dubbed CalCompute, intended to support research under California's regulatory framework.
We're fighting for a national AI strategy that gives Little Tech a fair shot and keeps the U.S. in the lead.
– Collin McCune (@Collin_McCune) September 29, 2025
CA's AI bill (SB 53) includes some thoughtful provisions that account for the distinct needs of startups. But it misses an important mark by regulating how the technology...
Industry reactions to SB 53 have been more restrained than during last year's debate. Anthropic openly supported the measure in its final stages, with co-founder Jack Clark stating that the law provides "real accountability" in AI development and deployment. OpenAI expressed encouragement that California had created a framework that could potentially align with federal regulations. Meta echoed a similar sentiment, describing the law as a constructive step.
However, not all responses were positive. The Chamber of Progress, a lobbying group, warned that the law could discourage startups and signal to younger companies that they might be better off building elsewhere. Venture capital firm Andreessen Horowitz noted that, while the bill included thoughtful provisions, its regulatory approach could slow innovation and further consolidate market power among the largest players.
At the federal level, efforts to establish a national AI framework remain unsettled. Some members of Congress have proposed federal standards that would allow states to set complementary rules, while others including Senator Ted Cruz argue that a state-by-state approach risks fracturing the national market and undermining US competitiveness with China.
Cruz has pledged to revive his push to prevent states from enacting their own AI regulations, calling a patchwork of standards "cataclysmic."