California Governor Gavin Newsom vetoed Senate Bill 1047 (SB 1047), a landmark piece of legislation that aimed to impose strict safety and security regulations on the development of advanced artificial intelligence (AI) systems. Introduced by Senator Scott Wiener, the “Safe and Secure Innovation for Frontier Artificial Intelligence Models Act” was intended to regulate powerful AI models, particularly those with the potential to pose significant risks to public safety.
The bill had garnered widespread support from AI researchers, including renowned figures such as Geoffrey Hinton and Yoshua Bengio, and had passed through the California Senate, making its way to Governor Newsom’s desk for final approval. However, in his veto message, the Governor outlined his concerns that the bill’s approach to AI regulation was overly narrow and could hinder innovation in the AI sector.
Governor Newsom acknowledged the importance of regulating AI systems but expressed doubts about the bill’s framework. “SB 1047 magnified the conversation about threats that could emerge from the deployment of AI,” Newsom wrote. However, he noted that the bill’s focus on the cost and scale of AI models—requiring stringent oversight for models that cost over $100 million to train—could give a “false sense of security” and overlook other potentially dangerous AI systems that might be smaller in scale but equally or more dangerous.
Newsom further emphasized the need for adaptability in AI regulation, pointing out that AI technology is still in its infancy, and solutions to manage its risks are rapidly evolving. He argued that any regulatory framework should be flexible enough to keep pace with technological advancements. “Adaptability is critical as we race to regulate a technology still in its infancy,” Newsom stated. He stressed that a regulatory framework should consider the actual risks posed by an AI system rather than focusing solely on the size and cost of the models.
Governor Newsom also raised concerns that SB 1047 would impose excessive regulation on even basic AI functions if they were part of larger systems. He argued that this approach could stifle innovation, particularly in California, which is home to 32 of the world’s 50 leading AI companies. “While well-intentioned, SB 1047 does not take into account whether an AI system is deployed in high-risk environments, involves critical decision-making, or the use of sensitive data,” Newsom explained. By applying stringent standards across the board, he believed the bill could curtail the very innovation that drives technological progress for the public good.
Despite vetoing SB 1047, Governor Newsom reiterated his commitment to regulating AI in a responsible and effective manner. He highlighted California’s ongoing efforts to manage AI risks, including actions taken under an Executive Order issued in September 2023, which directed state agencies to assess the potential threats and vulnerabilities posed by AI to California’s critical infrastructure. Newsom also noted that he had signed over a dozen bills regulating specific AI-related risks within the last 30 days.
Newsom pointed to federal efforts, including those led by the U.S. AI Safety Institute under the National Institute of Standards and Technology (NIST), as important initiatives informing AI risk management practices. He expressed his willingness to continue working with the Legislature, federal partners, and AI experts to develop a more comprehensive and science-based regulatory framework for AI.
SB 1047 was widely regarded as a pioneering effort to impose oversight on AI models with significant computational requirements. Supporters of the bill praised its focus on ensuring public safety while maintaining California’s leadership in AI innovation. Senator Wiener, the bill’s author, had argued that it was crucial for California to take proactive steps to regulate AI, given the state’s position as a global hub for AI development.
However, critics, including former Speaker of the House Nancy Pelosi, had expressed concerns that the bill did not fully address the complexities of AI regulation. Pelosi described the bill as “well-intentioned but ill-informed,” emphasizing the need for a more nuanced approach.
Need Help?
If you’re wondering how California’s AI laws and regulations, or any other AI legislation around the world, could impact you, don’t hesitate to reach out to BABL AI. Their Audit Experts are ready to provide valuable assistance while answering your questions and concern