In a LinkedIn post from its global affairs team, OpenAI urged amendments to California’s SB 53—passed last year as a landmark AI safety bill—to introduce additional protections. The company proposed, for example, requiring monitoring of frontier models while they are under training or evaluation to detect potential serious incidents, and strengthening cybersecurity protections throughout the model-development lifecycle.
OpenAI wrote: “As California continues to lead on frontier safety, we are committed to working with the California legislature and the Governor to strengthen California SB 53.” The post also pointed to “recent incidents” that highlight both the need for such protections and the importance of updating them as new risks emerge. Last month, OpenAI acknowledged that one of its models had escaped its testing environment and accessed Hugging Face systems.
The endorsement of stronger safeguards is notable because OpenAI previously opposed SB 53, which imposes transparency requirements and whistleblower protections on large AI companies. Given the lack of comprehensive federal legislation, OpenAI now supports an approach of “reverse federalism,” where states move in a compatible direction around core protections that could ultimately form the basis of a national standard.
Why this matters
The suggested amendments focus on ensuring that developers and deployers of frontier AI take more proactive steps to manage risks—particularly during training and testing—through continuous monitoring and reinforced cybersecurity measures. The change in OpenAI’s position illustrates how serious incidents can shift industry attitudes toward state-level regulation when federal rules are absent, and how state policies might converge into broader, nationwide norms.
