Scott Wiener, Author of California's AI Safety Bills, Takes Pelosi's Seat Seeking Federal Rules

Wiener set out his federal agenda in an interview with MIT Technology Review conducted on Friday, a few hours after he was declared the winner. He told the publication that he had hoped Congress would act decisively on AI before the end of this year, but that "that does not appear to be happening." Democrats have discussed forming a select committee on artificial intelligence since September, and Wiener said he would be keen to sit on it.

His election moves one of the AI safety movement's most effective legislators into the body that would write national rules. His 2024 bill SB 1047 was set to become the country's first substantial frontier AI law before Governor Gavin Newsom vetoed it. He then sponsored SB 53, a lighter-touch measure that Newsom signed. Both bills have inspired proposed federal legislation.

Wiener defeated San Francisco Supervisor Connie Chan in California's 11th District, which covers the whole city. Chan, whom Pelosi endorsed in May, conceded the day after the election, saying she did not have a "clear path to victory," according to KTVU. Early returns gave Wiener about 55% of the vote to Chan's 44%.

In the interview, Wiener said he wants "a strong, empowered, nimble federal regulator," pointing to legislation proposed by Representative Ro Khanna, a fellow California Democrat, to create one. He also backed whistleblower protections, testing and auditing requirements and a kill-switch requirement for AI systems, and said Congress should use its oversight powers to hold AI labs to account.

He is opposed to broad federal preemption of state AI laws, which some of the federal proposals would impose. "States are, as Brandeis says, laboratories," he said, adding that states can move faster than Congress.

Wiener framed his urgency around a report released the day before the interview by METR, an AI evaluation nonprofit, which MIT Technology Review said had been widely read as evidence that Anthropic had achieved recursive self-improvement, in which AI meaningfully contributes to its own development. "It's very concerning," Wiener said. "We need to slow this down. Because it's moving very quickly, and we could sleepwalk into a disaster."

His campaign received more than $3 million from Public First Action, a nonprofit that favours AI regulation and receives substantial funding from Anthropic. Opponents used that support to argue he was too close to the AI industry. Wiener rejected the charge, writing after the interview that the attacks were "shockingly cynical" and that Public First backed him because he was the first lawmaker in the country to introduce and pass AI regulation.

The national politics of AI may not match his priorities. Voters approved 40 of 41 ballot measures opposing data centres in last Tuesday's elections, but Wiener said he sees that revolt as driven mainly by local concerns such as electricity bills, water use and the effect on nearby homes, rather than by wider fears about where AI is heading.

He described himself as only "cautiously optimistic" about Congress's ability to rein in the technology, given its record on regulating the tech industry. He also expects difficulty getting legislation past President Donald Trump, who called concerns about catastrophic AI risk a "hoax" in September, and he views cooperation with China, which Trump has resisted, as essential to avoiding the worst outcomes.

"So far what we've seen are some problematic things, but not complete disasters," Wiener said of the administration. "I hope that it doesn't take a disaster to spur the president into acting."