Tech Entrepreneur Warns Calif. Against ‘Awful’ AI Regulation
California is considering some “awful” regulations for AI, tech entrepreneur Andrew Ng said Thursday. His comments came a day after California Gov. Gavin Newsom (D) warned state lawmakers against overregulating AI.
Speaking at an AI workshop co-hosted by DOJ, Ng criticized California’s SB-1047. The bill would let the state attorney general take legal action against large AI developers if they cause “severe harm” to state residents. Sen. Scott Wiener (D) introduced the bill.
It’s the equivalent of holding a motor company liable for harm caused by anyone who uses the motor for “nefarious purposes,” said Ng, former head of Google Brain and chief scientist at Baidu, now an adjunct professor at Stanford University. “It would be awful for American innovation and for California. I can’t believe that California, home to such innovation, is contemplating such a ridiculous regulation.”
The legislation requires developers of large AI models to provide transparent pricing to “ensure startup developers have equal opportunities to compete and innovate.” The Senate voted 32-1 to pass the bill on May 21. The California Assembly has until Aug. 31 to take up the legislation.
Newsom said, “I want [California] to continue to dominate" the tech sector, during a University of California, Berkeley panel discussion Wednesday evening: “If we over-regulate, if we overindulge, if we chase the shiny object, we could put ourselves in a perilous position.” He said there’s a risk of tunnel vision given the “regulatory mindset” not to make the same mistakes that were made with social media.
DOJ Antitrust Division Chief Jonathan Kanter and European Commission President Vera Jourova spoke at Thursday’s workshop that Stanford co-hosted. They said enforcers should act in ways that allow newcomers to challenge dominant players like Google, Meta and Microsoft.
Kanter highlighted historic antitrust cases against Standard Oil, AT&T and Microsoft, which he said let nascent companies compete. AI presents a similar threat of market concentration and, given AI’s reliance on data and computing power, tech companies could gain even more influence over emerging technology, he said.
Indeed, it’s hard to see startups challenging Google, Meta, Microsoft and OpenAI, given the infrastructure required to collect and store data, Jourova said. California in particular, with its regulation and business innovation, will have a major impact in Europe and across the world, she said: There’s a “narrow opportunity” right now to set proper rules for AI.
She discussed how AI technology is making it difficult for social media users to discern if content is real or synthetic. Elections must remain fair with open debate between “real human beings,” she said: Campaigns shouldn’t become a competition between candidates using “manipulative” techniques. She noted how Russia, China and Iran are using digital tools that spread disinformation and interfere in elections. “The essence of democracy is trust,” she said. “Without trust, democracy crumbles.”
Newsom said some constituents are concerned the 2024 election will be the last “free and fair” contest due to the increasing role of deepfake videos. He signed a law in 2019 banning the use deepfakes to influence political campaigns after the emergence of an AI-generated video purporting to show then-House Speaker Nancy Pelosi, D-Calif., drunk. “I’ve got personal reasons to believe [concerns about deepfakes are] legit: the voice, the videos, the AI bots, the persuasion.”