Recently, California Governor Gavin Newsom vetoed the highly anticipated SB1047 bill on Monday, which aimed to regulate the development of large-scale artificial intelligence models. This decision has sparked extensive discussions in both the tech and political spheres.
SB1047, proposed by State Senator Scott Wiener, required companies developing AI models with a cost of at least $100 million and using 10^26 FLOPS of computational power to implement safety protocols. The bill passed through the California legislature but faced opposition from Silicon Valley heavyweights, including OpenAI and Meta's Chief AI Scientist Yann LeCun.
In his veto statement, Newsom said, "SB1047 has good intentions, but it fails to consider the specific application environments and risk levels of AI systems, instead imposing strict standards on all large systems. This is not the best way to protect the public."
Image Source: The image was generated by AI, provided by the image licensing service Midjourney
Former Speaker of the House Nancy Pelosi also criticized the bill as "well-intentioned but lacking in foundation," agreeing with Newsom's decision. Meanwhile, Newsom emphasized that he has signed 17 AI-related bills and invited experts like Fei-Fei Li to assist in formulating viable AI regulatory measures.
Bill proposer Wiener stated that this veto is a "setback for those who believe in overseeing large corporations," but believes that the related debate has advanced the international discourse on AI safety.
This decision underscores the complexity of AI regulation and the challenges of balancing innovation with ensuring safety. As AI technology continues to rapidly evolve, effective regulation will remain a critical issue for policymakers.