Sacramento, CA – The California State Senate granted final approval early Saturday morning to Senate Bill 53 (SB 53), a significant artificial intelligence safety measure that establishes new transparency mandates for large AI development companies. The legislation now proceeds to Governor Gavin Newsom for signature or veto, marking a critical juncture for AI regulation within the state.
Authored by State Senator Scott Wiener, SB 53 outlines key provisions: it "requires large AI labs to be transparent about their safety protocols, creates whistleblower protections for [employees] at AI labs & creates a public cloud to expand compute access (CalCompute)," according to Wiener. These requirements could impact firms developing and deploying advanced AI models across various sectors, including advanced manufacturing, supply chain optimization, and logistics.
Governor Newsom has not publicly commented on SB 53. However, last year, he vetoed a more expansive AI safety bill, also authored by Senator Wiener, while endorsing narrower legislation primarily targeting issues such as deepfakes. At that time, Governor Newsom cited concerns that the previous bill applied "stringent standards" to large models irrespective of their deployment in "high-risk environments, [involving] critical decision-making or the use of sensitive data." Senator Wiener has indicated that the current iteration of SB 53 was influenced by recommendations from an AI expert panel convened by Governor Newsom after the prior veto.
Recent amendments to SB 53 introduce a tiered disclosure system. Companies developing "frontier" AI models with less than $500 million in annual revenue will be required to disclose high-level safety details. Conversely, companies exceeding $500 million in annual revenue will be mandated to provide more detailed reports.
The bill has encountered opposition from several Silicon Valley companies, venture capital firms, and lobbying organizations. OpenAI, in a recent letter to Governor Newsom, did not specifically mention SB 53 but advocated for companies to be considered compliant with statewide safety rules if they meet federal or European standards, thereby avoiding "duplication and inconsistencies." Additionally, Andreessen Horowitz's head of AI policy and chief legal officer has claimed that "many of today’s state AI bills — like proposals in California and New York — risk" violating constitutional limits on states' authority to regulate interstate commerce.
In contrast, Anthropic has publicly endorsed SB 53. Jack Clark, co-founder of Anthropic, stated that while the company would prefer a federal standard, "in the absence of that this creates a solid blueprint for AI governance that cannot be ignored."
