California's controversial AI safety bill, SB-1047, has passed the State Assembly and is now one step away from Governor Gavin Newsom's desk. Newsom must decide by September 30 whether to sign it into law.

California Governor Gavin Newsom has vetoed SB 1047, a significant bill aimed at regulating artificial intelligence development. Authored by State Senator Scott Wiener, the bill sought to impose liability on companies developing AI models, requiring them to implement safety protocols to mitigate potential "critical harms." The regulations would have specifically targeted models with a training cost of at least $100 million and those utilizing 10^26 FLOPS (floating point operations), a measure of computational power. The bill faced considerable opposition from various stakeholders in Silicon Valley, including prominent companies like OpenAI and influential technologists such as Yann LeCun, Meta's chief AI scientist. Even some Democratic politicians, including U.S. Congressman Ro Khanna, expressed their disapproval. Despite amendments made to the bill in response to feedback from AI companies like Anthropic, concerns remained about its implications. Newsom had previously indicated his reservations about the bill, and upon announcing the veto, he articulated that while the bill was well-intentioned, it failed to consider the context in which AI systems are deployed, particularly in high-risk environments or when sensitive data is involved. He criticized the bill for applying stringent standards to basic functions of large systems, arguing that this approach would not effectively protect the public from genuine threats posed by AI technology. Nancy Pelosi, a long-time Congresswoman and former House Speaker, also criticized SB 1047, labeling it as "well-intentioned but ill-informed." Following the veto, she commended Newsom for recognizing the need to empower smaller entrepreneurs and academic institutions rather than allowing large tech companies to dominate the field. In conjunction with the veto, Newsom's office highlighted his recent efforts to regulate AI technology, noting that he had signed 17 bills related to AI regulation in the past month. He has sought the expertise of notable figures in the field, such as Fei-Fei Li, Tino Cuéllar, and Jennifer Tour Chayes, to help California establish effective guidelines for the deployment of generative AI. Fei-Fei Li, often referred to as the "godmother of AI," had previously warned that SB 1047 could harm California's emerging AI ecosystem. In response to the veto, Senator Wiener expressed disappointment, describing it as a setback for those advocating for oversight of large corporations that make critical decisions impacting public safety and welfare. He asserted that the discussions surrounding the bill had significantly advanced the global conversation on AI safety.

California Governor Gavin Newsom has vetoed SB 1047, a significant bill aimed at regulating artificial intelligence development. Authored by State Senator Scott Wiener, the bill sought to impose liability on companies developing AI models, requiring them to implement safety protocols to mitigate potential "critical harms." The regulations would have specifically targeted models with a development cost of at least $100 million and those utilizing a substantial amount of computational power during training. The bill faced considerable opposition from various stakeholders in Silicon Valley, including prominent companies like OpenAI and influential technologists such as Yann LeCun, Meta's chief AI scientist. Even some Democratic politicians, including U.S. Congressman Ro Khanna, expressed concerns about the bill. Despite amendments made to address some of these concerns, the opposition remained hopeful that Newsom would ultimately veto the legislation, especially given his previous indications of reservations about it. In his veto statement, Newsom articulated that while the intentions behind SB 1047 were commendable, the bill failed to consider the context in which AI systems are deployed, particularly in high-risk environments or when sensitive data is involved. He criticized the bill for applying stringent standards to basic functions of AI systems, arguing that this approach would not effectively protect the public from genuine threats posed by the technology. Nancy Pelosi, a long-time Congresswoman and former House Speaker, also criticized the bill, labeling it as "well-intentioned but ill-informed." Following the veto, she commended Newsom for recognizing the need to empower smaller entrepreneurs and academic institutions rather than allowing large tech companies to dominate the field. In conjunction with the veto, Newsom's office highlighted his recent efforts to regulate AI technology, noting that he had signed 17 bills related to AI regulation in the past month. He has also sought guidance from experts in the field to help California establish effective frameworks for the responsible deployment of generative AI. In response to the veto, Senator Wiener expressed disappointment, describing it as a setback for those advocating for oversight of large corporations that make critical decisions impacting public safety and welfare. He emphasized that the discussions surrounding the bill have significantly advanced the conversation about AI safety on a global scale.