California Senate passes AI safety bill, now awaits Governor's approval
What's the story
The California State Senate has passed a major artificial intelligence (AI) safety bill, SB 53. The legislation was approved in the early hours of Saturday and now awaits Governor Gavin Newsom's signature or veto. The bill mandates large AI labs to disclose their safety protocols and provides whistleblower protections for employees working in these labs. It also proposes a public cloud platform, CalCompute, to broaden compute access.
Bill details
SB 53 will make AI labs more accountable
The bill's author, State Senator Scott Wiener, said that SB 53 will make AI labs more accountable. It will also protect whistleblowers who expose safety concerns within these organizations. The public cloud platform CalCompute is expected to democratize access to computing resources. This move could potentially level the playing field between large tech companies and smaller start-ups or researchers.
Governor's position
Newsom previously vetoed a similar bill
Governor Newsom has not publicly commented on SB 53. However, last year, he vetoed a more comprehensive safety bill by Wiener while approving narrower legislation addressing issues such as deepfakes. At the time, Newsom had stressed the need to "protect the public from real threats posed by this technology." However, he had also criticized Wiener's earlier bill for imposing "stringent standards" on large models irrespective of their deployment in high-risk environments or involvement in critical decision-making processes.
Bill amendments
SB 53 has faced opposition from Silicon Valley firms
The latest version of SB 53 requires companies developing "frontier" AI models with annual revenues under $500 million to disclose only high-level safety details. Companies making over this amount will have to provide more detailed reports. The bill has faced opposition from several Silicon Valley firms, venture capitalists, and lobbying groups. OpenAI recently argued in letter to Newsom that companies should be considered compliant with statewide safety rules if they meet federal or European standards, without specifically mentioning SB 53.
Regulatory debates
Andreessen Horowitz's backing for Trump raises eyebrows
Andreessen Horowitz (a16z)'s head of AI policy and chief legal officer recently claimed that many state AI bills, including those in California and New York, risk violating constitutional limits on how states can regulate interstate commerce. The firm's co-founders had previously cited tech regulation as a reason for supporting Donald Trump's second presidential bid. Meanwhile, Anthropic has endorsed SB 53 with co-founder Jack Clark saying it provides "a solid blueprint for AI governance that cannot be ignored."