Ahead of the presidential debate between U.S. Vice President and Democratic presidential candidate Kamala Harris and former U.S. President and Republican presidential candidate Donald Trump on September 10, 2024 at the National Constitution Center in Philadelphia, Pennsylvania. , California Governor Gavin Newsom speaks to the press.
Matthew Hatcher | AFP | Getty Images
California Governor Gavin Newsom on Sunday vetoed a hotly contested artificial intelligence safety bill after the tech industry voiced opposition to the bill, saying it could drive AI companies out of the state and stifle innovation. was activated.
Newsom said the bill “does not take into account whether AI systems are being deployed in high-risk environments, whether they involve critical decision-making or the use of sensitive data,” adding that “the most basic functions “Strict standards will also be applied.” A large system will deploy it. ”
Newsom asked leading experts in generative AI to help California “develop actionable guardrails” focused on “developing empirical, science-based trajectory analysis.” said. He also ordered state agencies to expand their assessment of potential catastrophic risks associated with the use of AI.
Generative AI that can create text, photos, and videos in response to free-form prompts could make some jobs obsolete, disrupt elections, and potentially overwhelm humans with devastating effects. This is causing both anxiety and excitement at the same time.
The bill’s author, Democratic Sen. Scott Wiener, said the bill is needed to protect the public before advances in AI get out of hand or out of control. California’s AI industry is growing rapidly, and some leaders have questioned the future of these companies in the state if the bill becomes law.
Wiener said Sunday that the veto would make California less safe and mean “companies looking to develop extremely powerful technologies will not be legally bound.” He added: “Voluntary efforts by industry are unenforceable and rarely work out well for the public.”
“We cannot wait for a major catastrophe to occur before taking action to protect the public,” Newsom said, adding, “This is not based on AI systems or empirical trajectory analysis of AI systems. “We have to compromise on solutions,” he added. ability. ”
Newsom said he will work with Congress on an AI bill in the next Congress. This comes as legislation to create safeguards stalls in the US Congress and the Biden administration moves forward with regulatory AI oversight proposals.
“A California-only approach may be warranted, especially in the absence of federal action by Congress,” Newsom said.
The Chamber of Progress, a tech industry coalition, praised Newsom’s veto, saying, “California’s high-tech economy has always thrived on competition and openness.”
Among other things, the measure would require safety testing for many cutting-edge AI models that cost more than $100 million to develop or require a specified amount of computing power. Developers of AI software operating in the state would also have been required to outline how to turn off their AI models, effectively a kill switch.
The bill would create a state agency to oversee the development of so-called “frontier models” that go beyond the capabilities of state-of-the-art existing models.
The bill faced strong opposition from a wide range of groups. Alphabet GOOGL.O’s Google, Microsoft MSFT.O-backed Open AI, and Meta Platforms META.O, which are developing generative AI models, had expressed concerns about the proposal.
Some Democrats in Congress, including Representative Nancy Pelosi, also objected. Supporters include Tesla Inc TSLA.O CEO Elon Musk, who also runs artificial intelligence (AI) company xAI. Amazon AMZN.O -backed Anthropic said the benefits of the bill would likely outweigh the costs, but added there were still concerns and ambiguities.
Separately, Newsom signed a bill requiring the state to assess the potential threat generative AI poses to California’s critical infrastructure.
Newsom said the state has analyzed energy infrastructure risks and previously convened power sector providers, and plans to conduct a similar risk assessment next year for water infrastructure providers and then for the communications sector. .
