Tesla CEO and xAI founder Elon Musk has thrown his support behind a California bill that’s divided Silicon Valley and promises to regulate the artificial intelligence industry.
The bill, SB 1047, would mandate companies that have spent more than $100 million to develop an AI model, or a model that reaches a certain threshold of computing power, to implement safeguards and conduct safety testing. The state attorney general would also be given the power to take action against developers whose model causes “severe harm.” Companies would also need to implement a kill switch to instantly terminate their models at any time.
“This is a tough call and will make some people upset, but, all things considered, I think California should probably pass the SB 1047 AI safety bill,” Musk wrote on X, formerly, Twitter, which he owns, on Monday. “For over 20 years, I have been an advocate for AI regulation, just as we regulate any product/technology that is a potential risk to the public.”
The bill passed the state’s Senate and is scheduled for a vote in the State Assembly later this week, before returning for a final vote in the Senate. Gov. Gavin Newsom would then have to sign the bill into law.
Although SB 1047 was proposed by Scott Weiner, a Democratic state senator representing San Francisco, it’s been opposed by a number of high profile Golden State Congressional Democrats, including Reps. Nancy Pelosi and Rho Khanna. Several tech companies have also expressed their opposition or proposed amendments to make SB 1047 more attractive.
Weiner eliminated earlier aspects of the bill, including criminal penalties for perjury and the creation of a government oversight committee dedicated to AI. He was advised, in part, by Anthropic, an AI-firm backed by Amazon and founded by ex-OpenAI engineer, which made a series of recommendations.
In a letter to Newsom earlier this month, Anthropic said SB 1047's “benefits likely outweigh its costs,” although it wasn’t completely certain. The Chamber of Progress, a progressive tech group that works with firms like Apple and Lyft, has removed its opposition to the bill, citing an amended provision on data transparency. SB 1047 is also supported by Geoffrey Hinton and Yoshua Bengio, researchers sometimes called the “godfathers of AI” and the Center for AI Safety.
On the other hand, ChatGPT maker OpenAI has opposed the bill, arguing that it would slow progress and push companies to innovate elsewhere. Google and Meta have expressed their concerns over the bill. as have a16z’s Marc Andreessen and Ben Horowitz. The third AI godfather and Meta’s chief AI scientist, Yann LeCun, has opposed SB 1047, as have a number of other tech firms.
“California should be leading the way on accelerating AI (safely), not creating the template to slow things down,” Box CEO Aaron Levie wrote Sunday. “We’re simply too early in the state of AI to taper progress.”