The vetoed legislation, SB 1047, aimed to impose stringent safety and transparency requirements on developers of large AI models, such as OpenAI’s GPT-4 and Meta’s LLaMA. Had it passed, the bill would have required companies to test their models rigorously and disclose safety measures publicly. The goal was to mitigate risks posed by AI, such as the potential for large models to be catastrophically misused in critical areas like energy infrastructure or bioweapon development.
However, Newsom rejected the bill, citing its overly broad application to all AI systems, regardless of risk level. He argued that the legislation could stifle innovation by applying the same regulatory standards to basic AI models as to high-risk, complex systems. Newsom suggested that California needs a more targeted approach to AI regulation—one that balances safety with the state’s leadership in AI development.
The bill faced significant opposition from tech giants and startups alike, many of whom warned it would hurt California’s thriving AI industry. Former US House Speaker Nancy Pelosi was among those who argued that the bill’s stringent requirements could discourage AI investment and innovation within the state.
Proponents of the bill, including figures like Elon Musk, the AI startup Anthropic, as well as some of Hollywood’s biggest names like Pedro Pascal and J.J. Abrams, stressed the importance of transparency and accountability in AI development, particularly as the technology rapidly advances without clear oversight.
In his statement, Newsom emphasized that California remains committed to shaping AI regulation and announced plans to collaborate with industry leaders to create more flexible guardrails. Experts, including AI pioneer Fei-Fei Li, will work with the state to develop a regulatory framework that fosters innovation while addressing safety concerns.
