California Gov. Gavin Newsom (D) on Sunday vetoed a landmark artificial intelligence (AI) bill that would create new safety rules for emerging technologies, a big win for much of Silicon Valley. brought about.
Mr. Newsom's veto is a major change in how he will act on the controversial bill known as California Senate Bill 104, or the Frontier Artificial Intelligence Models Safe and Secure Innovation Act. It ended weeks of skepticism.
in Veto message In the paper published Sunday, the governor said the bill's focus on the “most expensive and large-scale models” could “give the public a false sense of security about control” of AI. said.
“Smaller, more specialized models could emerge that are just as dangerous, if not more dangerous, than those covered by SB 1074. At the expense of prioritizing the public interest and promoting progress,” he said, “Innovation itself may be suppressed.”
The bill passed the state House before being sent to his desk late last month, and his veto came just a day before Monday's deadline.
The bill, also known as SB 1047, would require powerful AI models to undergo safety testing before being released to the public. This could include, for example, testing whether a model can be manipulated to hack the state's power grid.
It was also intended to hold developers accountable for serious harm caused by their models, but it only applied to AI systems that cost more than $100 million to train. There is no current modelStill reaching that number.
“SB 1047, while well-intentioned, fails to consider whether AI systems are deployed in high-risk environments, involve critical decision-making, or involve the use of sensitive data,” Newsom said. he wrote. “Instead, this bill applies strict standards to even the most basic functions, so long as large systems implement them.”
Mr. Newsom has often expressed skepticism about reining in AI technologies that bring billions of dollars to the Golden State. According to Newsom's office, California is home to 32 of the world's 50 largest AI companies, making it a major center for AI law.
The governor stressed that his veto does not mean he disagrees with the author's argument that we urgently need to respond to advancing technology to prevent catastrophe.
“California is not going to abdicate its responsibility,” he said, adding, “aggressive guardrails must be in place and serious consequences for bad actors must be clear and enforceable.” Ta.
Newsom argued that any solution must be informed by “empirical trajectory analysis” of AI systems and their capabilities.
The bill received mixed opinions from AI startups, big tech companies, researchers, and even some members of Congress who were divided on whether to curb technology development or establish much-needed guardrails. It was done.
People on both sides of the debate have ramped up pressure on Newsom over the past few months.
Some of the nation's largest technology companies, including open AI, Google and meta Facebook and Instagram's parent companies expressed concern that the bill would target AI developers rather than abusers, and argued that technology safety regulations should be determined at the federal level.
Meanwhile, leading AI startup Anthropic I mentioned merit last month. The bill probably outweighed the risk.
last week, 120+ Hollywood Figures “The most powerful AI models could soon pose significant risks,” he wrote in an open letter urging the bill to be signed. Earlier this month, more than 100 current and former employees of major AI companies, including OpenAI, Anthropic, Google's DeepMind, and Meta, also sent a letter to Newsom warning of similar risks.
Members of Congress also participated in the debate, including former Speaker Nancy Pelosi (D-Calif.) and other California politicians who oppose the bill. Pelosi said last month that “many” in Congress view the bill as “well-intentioned but poorly informed.”
Mr. Newsom pushed back on arguments that California should not play a role in legislation that affects the entire country.
He pointed to the risk analysis currently being done on a federal and state basis regarding AI, saying, “A California-only approach could well be justified, especially in the absence of federal action by Congress. But it has to be based on empirical evidence and science.”
The governor signed a series of other bills earlier this month aimed at preventing the misuse of AI and putting guardrails around emerging technologies.
Three of these bills aim to prevent the misuse of sexually explicit deepfakes, which generate images, audio, and video and digitally alter likenesses and voices. He signed two other bills aimed at protecting the names, images and likenesses of actors and performers from being copied by artificial intelligence without their permission.
