In summary
The bill would require technology companies to test whether AI harms society. This drew opposition from many members of Congress and major AI companies such as Google, Meta, and OpenAI.
California Governor Gavin Newsom today vetoed the most ambitious and controversial bill to regulate artificial intelligence approved by Congress this year.
The bill, Senate Bill 1047, would test AI models to determine whether they have the potential to cause mass death, compromise public infrastructure, or enable serious cyberattacks. It was mandatory.
In his veto letter, Newsom criticized the bill, saying it could “stifle the very innovation that promotes progress in the public interest.” He also said it unnecessarily regulates AI used in low-risk situations and was created without sufficient research.
“While a California-only approach may be well justified, especially in the absence of federal action by Congress, it must be based on empirical evidence and science,” he wrote.
The bill would apply only to the most expensive AI models, which require more than $100 million to develop, and Newsom opposed that standard, saying even cheaper technology could still be harmful.
The governor wrote that he will work to “find an appropriate path forward, including through legislation and regulation” to address AI risks.
Opponents argued that the bill would harm the state’s economy and the AI industry. They included Google, Meta, OpenAI, and eight members of California’s congressional delegation. The state Chamber of Commerce praised the veto, saying the bill “puts California’s status as a global hub of innovation at great risk.”
The bill’s sponsor, Sen. Scott Wiener, D-San Francisco, said the veto was a missed opportunity for “California to once again lead the way in innovative technology regulation” and protect public safety. According to one poll, support for the bill includes 59% of California voters, as well as billionaire Elon Musk, the Screen Actors Guild, the Service Employees International Union, the National Organization for Women, and people who work at AI manufacturing companies. This includes whistleblowers who were
One supporter, Teri Ole of the advocacy group Economic Security California, said the governor’s veto would mean “we lose the opportunity to lead.”
Newsom signed about a dozen other AI-related bills this month, including bills to protect voters from deepfakes and creators from unauthorized digital replicas of their likenesses. He also signed legislation that would require companies to share information about how they train their generative AI models and allow users to determine whether media was created by AI.
Overall, the California Legislature passed more than 20 bills this year to regulate artificial intelligence.
Speaking at the Generative AI Symposium in May, Newsom said it was important to respond to calls for oversight by some AI developers, but he also said he didn’t want to overregulate a critical industry. I warned you. According to Forbes, California is home to 35 of the top 50 AI companies, and according to Crunchbase, Silicon Valley-based companies receive more AI investment than any other region. .
People in the industry disagree on SB 1047, and some startups like Anthropic have joined whistleblowers in supporting SB 1047. This reflects this year’s debate over whether to ban the weaponization of robots. A bill to do so was co-sponsored by Boston Dynamics, a major robot manufacturer. Despite that support, Newsom vetoed the bill, arguing that police need immunity.
Newsom’s veto of SB 1047 prevents California from aligning its AI regulations with those of the European Union. Representatives from the European Union told CalMatters earlier this year that the union’s AI law, said to be the most comprehensive regulation, would require states to include SB 1047 and other measures such as watermarking AI-generated images. He said that if the law were enacted, it could almost be replicated in California. To protect people from automated discrimination. The latter two bills failed to pass Congress.
Regardless of the governor’s decision, the debate over the Wiener bill has brought many people into AI policy who don’t normally participate in the discussion, said Alondra Nelson, former White House director of science and technology policy.
Nelson was disappointed that the bill did not address the threat to civil rights posed by AI or the need to protect people from it in the workplace. But she liked the testing requirements. That was consistent with the provisions of the White House’s AI Bill of Rights blueprint, which Nelson helped write.
Nelson said he hopes the coalition formed to support SB 1047 will help advance future AI legislation. San Ramon Democratic Rep. Rebecca Bauer-Kahan, who worked with Nelson on an unsuccessful bill to protect against AI-based discrimination, said the AI Civil Rights Act, which includes similar provisions, was introduced in Congress this week. He said he plans to reintroduce a similar bill next year. . She also expects work to continue passing regulations like SB 1047.
“California is one of the few places in the United States where we still demonstrate the ability and willingness to govern technology,” she said.
September 6, 2024September 6, 2024
