California state legislators just took the nation’s most significant step yet toward regulating artificial intelligence. We are indeed pleased that they’ve recently passed final approval to SB 53, their big AI safety bill. Sponsoring legislator State Senator Scott Wiener tweeted that the bill creates new transparency requirements for the big corporations developing AI. The bill now heads to Governor Gavin Newsom to be signed into law. He has been averse to regulating the nascent AI sector.
SB 53 received a boost from a recent veto by Governor Newsom of a more expansive safety bill, SB 1047. This bill was authored by Wiener. To help address that veto, Newsom convened his own task force of AI experts. In addition, they provided concrete recommendations and examples for strengthening the current state of AI governance. “In the absence of that this creates a solid blueprint for AI governance that cannot be ignored,” said Jack Clark, who has been vocal about the need for regulatory standards.
The recently passed legislation makes a distinction among companies based on their annual revenue. Firms developing “frontier” AI models earning less than $500 million annually will only be required to disclose high-level safety details. By comparison, companies with more than $1 million in gross revenues are required to file much more detailed reports on their safety enforcement activities. This tiered approach increases accountability for larger entities and holds them to stricter standards. It gives less room to maneuver for smaller companies.
Not only does SB 53 increase transparency and accountability in AI technologies, it protects whistleblowers within AI laboratories. Alongside that work, it helped introduce CalCompute, a public cloud resource specifically meant to improve accessibility. “It requires large AI labs to be transparent about their safety protocols, creates whistleblower protections for employees at AI labs and creates a public cloud to expand compute access,” Wiener elaborated.
The bill has not been without its critics. Andreessen Horowitz’s head of AI policy and chief legal officer warned against state AI bills. In particular, they argued that many of these proposed measures in California would run afoul of constitutional prohibitions on restricting interstate commerce. The Trump administration and its pro-innovation allies are pushing very hard right now for a ten-year moratorium on any state-level AI regulation. They’re concerned that these types of measures will discourage innovation and expansion in the tech sector.
The AI regulation debate continues to grow. With major interests now applying pressure from all sides, Governor Newsom faces a difficult decision in determining whether SB 53 should become law or not. His decision will fundamentally change how California will govern AI. Beyond helping patients and providers in Arkansas, it would create an important policy precedent for other states passing similar legislation.
