Global: Newsom Vetoes Controversial California AI Bill

California Governor Gavin Newsom has vetoed a bill that would have introduced strict safety testing requirements for artificial intelligence (AI) companies developing models costing more than $100 million or those utilizing substantial computing power. The bill, known as SB 1047, also sought to mandate the inclusion of fail-safe mechanisms, or a “kill switch,” allowing developers to shut down AI models in the event of emergencies or unforeseen risks.

In his veto message on Sunday (Sept. 29), Newsom explained his concerns with the bill’s approach, stating that it focused too narrowly on the most expensive and large-scale AI models. He argued this could create a false sense of security for the public, as smaller, more specialized AI models could emerge as equally or even more dangerous than those targeted by SB 1047.

“By focusing only on the most expensive and large-scale models, SB 1047 establishes a regulatory framework that could give the public a false sense of security about controlling this fast-moving technology,” Newsom wrote in his correspondence to lawmakers. He further noted that the bill’s blanket approach risked stifling innovation, which he said is key to ensuring the benefits of AI are maximized for the public good.

While Newsom acknowledged the need for public safety measures regarding AI, he criticized the bill’s broad and inflexible framework. He advocated for AI regulation that is based on empirical evidence, tailored to specific risks associated with different AI applications, and capable of adapting to the rapidly evolving nature of the technology.

The Governor also highlighted California’s ongoing efforts to address AI risks through other avenues, including executive orders and recently passed legislation. He expressed his commitment to working with stakeholders across industries to create more targeted and scientifically informed AI regulations moving forward.

“California is home to 32 of the world’s 50 leading AI companies, pioneers in one of the most significant technological advances in modern history,” Newsom emphasized. “We lead in this space because of our research institutions, diverse workforce, and a culture that fosters intellectual freedom. I take seriously the responsibility to regulate this industry.”

The decision represents a victory for many AI companies, though the bill had strong backing from over 100 AI employees, including notable figures from OpenAI, Google DeepMind, Anthropic, Meta, and xAI. Among the signatories were Turing Award winner Geoffrey Hinton and University of Texas professor Scott Aaronson, who had urged Newsom to sign the bill, citing the potential risks posed by powerful AI models.

In a statement issued on Sept. 9, the employees expressed concerns about AI’s potential to create severe risks, including facilitating access to biological weapons and enabling cyberattacks on critical infrastructure. They argued that it was both feasible and necessary for frontier AI companies to implement reasonable safeguards and test their models for potentially catastrophic harm.

While the veto ends this particular legislative effort, Newsom’s stance suggests that AI regulation remains a top priority for California as the state seeks to balance innovation with safety in the fast-evolving field of artificial intelligence.