Stay up to this point with free updates
Simply log in Artificial intelligence myFT Digest – delivered straight to your inbox.
California Gov. Gavin Newsom has vetoed a controversial attempt to manage artificial intelligence, raising concerns that the bill could stifle innovation as a result of intense pressure from tech firms.
Newsom, a Democrat, waited until the eleventh hour to announce his decision after the bill passed the state legislature in late August.
The bill would have forced those developing essentially the most powerful AI models to follow strict rules, including implementing a kill switch to forestall catastrophic damage.
Leading AI firms, including Google, OpenAI and Meta, opposed the bill and lobbied heavily. They complained that premature laws could stifle AI development and threaten California's leading role in developing the technology. Amazon-backed Anthropic and Elon Musk, which owns the startup xAI, supported the laws.
In a letter to the state Senate on Sunday, Newsom defended his veto of the Safe and Secure Innovation for Frontier Artificial Intelligence Systems Act, generally known as SB 1047.
He said the framework could “limit the very innovations that drive progress for the common good,” noting that California is home to 32 of the world's leading AI firms.
In particular, he said targeting models by size — the bill would require safety testing and other guardrails for models that cost greater than $100 million to develop — was the incorrect benchmark. It could “give the general public a false sense of security about controlling this rapidly evolving technology” when “smaller, specialized models could prove equally or more dangerous.”
Senator Scott Wiener, who introduced the bill, said the “veto is a setback for anyone who believes in oversight of enormous corporations that make critical decisions that affect the protection and well-being of the general public and the long run of the planet.” impact”.
He added that firms looking for to develop extremely powerful technology face no binding restrictions from U.S. policymakers, especially given Congress's continued paralysis over regulating the tech industry in any meaningful way.
However, Newsom insisted that the bill didn’t keep in mind whether an AI system “is utilized in high-risk environments, requires critical decisions, or uses sensitive data.”
“Instead, the bill applies strict standards for even essentially the most basic functions – so long as they’re delivered across a big system. I don’t imagine that is one of the best approach to protecting the general public from real threats from technology.”
Over the past 30 days, Newsom has signed bills that cover the use and regulation of generative AI technology — the sort that creates text or images — including deepfakes, AI watermarking and misinformation.
Technology experts have also partnered with the state to assist develop “workable guidelines” for the usage of generative AI which might be supported by empirical and scientific evidence, he said.
The Artificial Intelligence Policy Institute, a think tank, called the governor's veto “misguided, reckless and antithetical to the people he’s alleged to govern.”
“Newsom had the chance to take a leadership role in regulating the democratic governance of AI development – a path he has taken in other industries – but has chosen to take our hands off the wheel, which can lead to this “This results in AI development getting off course in an uncontrolled manner,” said Managing Director Daniel Colson.
“Newsom and lawmakers must return to Sacramento next session to achieve agreement on a set of measures that set meaningful guardrails for AI development.”
