A controversial bill aimed at enforcing safety standards for large artificial intelligence models has now passed the California State Assembly by a 45-11 vote. Following a 32-1 state Senate vote in May, SB-1047 now faces just one more procedural state senate vote before heading to Governor Gavin Newsom’s desk.
SB-1047 asks AI model creators to implement a “kill switch” that can be activated if a model introduces “novel threats to public safety and security,” especially when acting with “limited human oversight, intervention, or supervision.” Critics argue the bill focuses on far-fetched future risks rather than present-day harms, such as deep fakes or misinformation.
In announcing the legislative passage, bill sponsor and state senator Scott Weiner cited support from AI industry luminaries like Geoffrey Hinton and Yoshua Bengio, who signed a statement warning of AI’s “risk of extinction.” Bengio contended in a recently published editorial in Fortune that the bill “outlines a bare minimum for effective regulation of frontier AI models.”
However, AI expert and Stanford professor Fei-Fei Li criticized the legislation, arguing in a separate Fortune editorial that the bill might result in developers acting defensively, impacting open-source sharing and academic research.
A group of California business leaders sent an open letter urging Newsom to veto the “fundamentally flawed” bill, citing concerns over regulatory ambiguity and potential stifling of innovation. Governor Newsom has until September 30 to decide, following which the legislature could override a veto with a two-thirds vote in each chamber.
Newsom previously expressed caution about over-regulation during a UC Berkeley Symposium in May, while also acknowledging the concerns raised by AI pioneers. “When you have the inventors of this technology saying, ‘Help, you need to regulate us,’ that’s a very different environment,” he noted.
For more details, read the full article on Ars Technica.