Anthropic endorses California’s AI safety bill, SB 53

In a significant move, Anthropic has publicly supported California's SB 53, a pioneering bill introduced by Senator Scott Wiener. This legislation aims to establish unprecedented transparency standards for the largest AI model developers globally. Anthropic's endorsement is particularly noteworthy given the opposition from major tech organizations like the Consumer Technology Association and the Chamber for Progress, which are actively lobbying against the bill. In a blog post, Anthropic expressed its belief that while the issue of frontier AI safety is better suited for federal oversight, the rapid advancements in AI technology necessitate immediate action rather than waiting for a consensus in Washington. They stated, "The question isn’t whether we need AI governance—it’s whether we’ll develop it thoughtfully today or reactively tomorrow. SB 53 offers a solid path toward the former." If enacted, SB 53 would require leading AI model developers such as OpenAI, Anthropic, Google, and xAI to establish safety frameworks and disclose public safety and security reports prior to launching advanced AI models. The legislation also proposes protections for whistleblowers who report safety violations. The bill specifically targets the prevention of "catastrophic risks," defined as events that could lead to at least 50 fatalities or over a billion dollars in damages. It notably focuses on preventing AI models from being utilized in the development of biological weapons or cyberattacks, rather than addressing more immediate concerns like deepfakes. The California Senate has already approved an earlier version of SB 53, but it requires a final vote before it can proceed to the governor. Governor Gavin Newsom has yet to comment on the bill, having previously vetoed a similar safety measure, SB 1047. Efforts to regulate frontier AI developers have met substantial resistance from both Silicon Valley and the Trump administration, which argue that such regulations could hinder American innovation in the global tech landscape. Notable investors, including Andreessen Horowitz and Y Combinator, have voiced concerns over SB 1047, while the Trump administration has indicated a willingness to block state-level AI regulations. Critics of state-led AI safety initiatives argue that these matters should be handled at the federal level. Matt Perault and Jai Ramaswamy from Andreessen Horowitz have raised constitutional questions regarding the impact of state laws on interstate commerce. Conversely, Anthropic co-founder Jack Clark has emphasized the urgency of establishing AI safety measures, regardless of federal action. He stated, "We have long said we would prefer a federal standard. But in the absence of that this creates a solid blueprint for AI governance that cannot be ignored." OpenAI's Chief Global Affairs Officer Chris Lehane previously urged Governor Newsom to avoid regulations that could drive startups out of California, though he didn’t specifically reference SB 53. Miles Brundage, a former OpenAI policy researcher, criticized Lehane’s remarks as misleading. Importantly, SB 53 specifically targets the largest AI firms, those generating over $500 million in revenue, and aims to enforce compliance with its safety requirements, imposing financial penalties for failures. Recently, California lawmakers modified SB 53 to eliminate a provision that would have mandated third-party audits for AI developers, a measure tech companies have resisted in previous discussions, citing the burdensome nature of such audits. Despite the backlash, experts believe that SB 53 represents a more measured approach to AI regulation than its predecessors, with hopes that it might soon become law.