California Governor Gavin Newsom on Sunday vetoed a bill that seeks to avoid catastrophic risks from advanced artificial intelligence models.
The bill, SB 1047, was the most controversial AI bill in Congress, with dozens of activists and AI companies siding with it on both sides. The authors warned that if left unregulated, AI models could be used to develop chemical and nuclear weapons, potentially causing mass casualties. Opponents argued that heavy-handed regulation would stifle AI development and force AI companies to leave the state.
In his veto message, Newsom said the bill addresses real issues but fails to establish an adequate regulatory framework.
“We do not believe this is the best approach to protecting the public from the real threats posed by technology,” he wrote. “Ultimately, any framework to effectively regulate aluminum will need to keep pace with the technology itself.”
Among the bill’s supporters was SAG-AFTRA, the union representing Hollywood actors. A group called “Artists for Safe AI” also issued an open letter on Tuesday in support of the bill, including JJ Abrams, Shonda Rhimes, Judd Apatow, Rob Reiner, Jane Fonda, Rian Johnson, Adam McKay, Mark Hamill, Mark Ruffalo, Don Cheadle, etc.
The Actors Guild has been outspoken about the threat of AI cloning actors and forcing them to appear in movies and TV shows without their consent. This is the first time the union has considered the broader risks that could arise from advanced AI models outside of the entertainment sector.
“This really stems from the fact that we have first-hand experience with the dangers of this aspect of AI,” said Jeffrey Bennett, the union’s general counsel. “This bill appears to be a bill that only targets incredibly powerful and expensive systems that can cause significant problems on a large scale. Why not regulate it at that level? Why not incorporate some sensible, basic safety protocols at this stage?”
In announcing his veto on Sunday, Newsom said he would continue to work on the issue next year, convening experts to develop regulations to encourage the safe development of AI.
SAG-AFTRA supported two other AI bills in California this year that would regulate the use of AI in entertainment. Newsom signed both bills at SAG-AFTRA headquarters earlier this month.
Labor unions were not involved in the discussion of SB 1047 while it was being considered in Congress. But on September 9, the union sent a letter to Newsom urging him to sign the bill.
“AI-powered deepfake technology is being used to create fake nude and pornographic images of SAG-AFTRA union members,” wrote Shane Guzman, a union lobbyist in Sacramento. “In our view, policymakers have a responsibility to intervene and protect their members and their citizens. SB 1047 is a prudent first step to get us there.”
Other unions and companies in Hollywood did not weigh in on the issue, only focusing on “frontier” AI models that currently don’t exist.
Newsom signed another AI bill, AB 2013, that would require AI developers to disclose whether they are training models based on copyrighted works. There is.
The Concept Art Association was a major supporter of the bill. This group represents artists who create visual ideas for movies, animation, and video games. In recent years, many people have seen their jobs captured and replaced by AI models.
The bill does not require developers to disclose the entire dataset used to train their models. You also don’t have to pay developers royalties for their copyrighted works. This issue is highly controversial and ongoing litigation. It simply requires developers to authorize the use of their copyrighted data and other “personal” information.
“Any disclosure we can get is a good thing,” said Deanna Igelsrud, a legislator and policy advocate at the Concept Art Association. “This is very common, but it’s a start.”
Hollywood unions are sponsoring a similar bill in Congress introduced by Rep. Adam Schiff earlier this year.
“These AI systems wouldn’t be able to output anything unless they were filled with all the art of world history,” Igelsrud says. “I don’t think people really understand that there’s a real human being attached to the data. Everyone assumes that if you put it on the internet, it’s free for everyone to use. That’s not the case. “
