California's state senate just passed SB 53, a scaled-back AI safety bill that zeros in on companies making over $500 million from AI models. Unlike last year's vetoed SB 1047, this narrower approach has Anthropic's backing and appears designed to survive Governor Gavin Newsom's scrutiny while still creating meaningful oversight of AI giants like OpenAI and Google DeepMind.
California's legislature just handed Governor Gavin Newsom what could be the state's first successful AI safety law. SB 53, authored by state senator Scott Wiener, cleared its final hurdle after learning harsh lessons from last year's regulatory defeat.
The new bill represents a strategic retreat from Wiener's previous attempt. Where SB 1047 cast a wide net that spooked the startup ecosystem, SB 53 laser-focuses on AI companies generating more than $500 million annually from their models. That threshold effectively targets the industry's power players - OpenAI, Google DeepMind, Meta, and Anthropic - while giving smaller companies breathing room.
The bill's requirements aren't revolutionary, but they're real. Major AI developers would need to publish safety reports for their models, immediately disclose any incidents to state authorities, and provide whistleblower channels for employees with safety concerns. For workers who've signed NDAs, that last provision could prove crucial as competition intensifies and safety protocols face internal pressure.
"We're entering this era where AI companies are becoming the most powerful companies in the world, and this is going to be potentially one of the few checks on their power," TechCrunch's Max Zeff said during a recent Equity podcast discussion about the legislation.
The timing matters as much as the content. Anthropic's endorsement signals that at least one major AI company sees regulatory cooperation as preferable to an adversarial relationship with lawmakers. That's a marked shift from last year's united industry opposition to SB 1047.
California's regulatory push comes as the federal landscape tilts decidedly pro-industry. The Trump administration has actively discouraged AI regulation, even inserting language into funding bills that would prevent states from implementing their own AI oversight. While those provisions haven't passed yet, they highlight the growing tension between federal deregulation and state-level safety concerns.
