California AI Safety Reports: SB 1047 Update
California’s AI Safety Law, SB 53, Returns to Test Tech Giants
Table of Contents
California is once again at the forefront of AI regulation with the potential passage of Senate Bill 53 (SB 53), a law requiring major AI developers to disclose details about their models’ capabilities and potential risks. This comes after a recent attempt in Congress to temporarily halt state-level AI laws failed, paving the way for continued state innovation in this rapidly evolving field.
The Push for AI Openness: What is SB 53?
SB 53, sponsored by State Senator Scott Wiener, aims to increase transparency around “high-risk” AI systems – those with significant capabilities that could pose risks to public safety. The bill builds upon previous, more ambitious proposals and focuses on compelling large AI developers to publish regular safety and security reports. These reports would detail how the AI models are tested, the potential harms they could cause, and the steps taken to mitigate those risks.
The legislation specifically targets companies developing frontier AI models – the most powerful and advanced AI systems currently available. It’s a direct response to concerns that the breakneck speed of AI advancement is outpacing efforts to understand and address potential dangers. The bill also echoes the principles outlined in the proposed federal RAISE Act, which similarly seeks to mandate safety reporting from large AI developers.
Federal Attempts to Preempt State Laws Fail
The future of state-level AI regulation was briefly uncertain when federal lawmakers considered a moratorium on state AI laws for up to ten years. the proposal, framed as a way to avoid a “patchwork” of regulations, faced widespread criticism from advocates who argued it would stifle innovation and leave the public vulnerable.Fortunately, a bipartisan coalition in the senate rejected the moratorium in a resounding 99-1 vote in July. This outcome signals strong support for allowing states to lead the way in establishing responsible AI practices.
“Ensuring AI is developed safely should not be controversial – it should be foundational,” stated Geoff Ralston, former president of Y Combinator.”Congress should be leading, demanding transparency and accountability from the companies building frontier models.But with no serious federal action in sight, states must step up. California’s SB 53 is a thoughtful, well-structured example of state leadership.”
A History of Resistance from Big Tech
Despite broad agreement on the need for transparency, getting AI companies to voluntarily comply with state-mandated requirements has proven challenging. Anthropic has publicly supported increased transparency and even expressed optimism about california’s AI policy recommendations. However, industry giants like OpenAI, Google, and Meta have been more hesitant.
This resistance is evident in recent decisions to delay or forgo publishing safety reports for their latest AI models. Google, for example, released Gemini 2.5 Pro without a corresponding safety report for months. OpenAI followed suit, launching GPT-4.1 without publicly detailing its safety assessments. A subsequent third-party study raised concerns that GPT-4.1 might be less aligned with human values than its predecessors, highlighting the potential risks of releasing powerful AI models without thorough evaluation.These actions underscore the importance of legislation like SB 53, which would legally obligate companies to provide crucial safety information. The lack of consistent reporting creates a significant information gap, making it tough for researchers, policymakers, and the public to assess the risks associated with these technologies.
What’s Next for SB 53?
SB 53 represents a compromise compared to earlier, more stringent AI safety bills. Still, it has the potential to considerably increase the amount of information available about the capabilities and risks of advanced AI systems.
Senator Wiener is once again pushing the bill forward, and the tech industry will be closely watching. The outcome of this legislative effort will not only shape the future of AI regulation in California but could also serve as a model for other states and potentially influence future federal policy. The debate surrounding SB 53 highlights the ongoing tension between fostering innovation and ensuring the responsible development and deployment of artificial intelligence.