OpenAI is urging California lawmakers to strengthen SB 53, the state’s frontier AI safety law. In a LinkedIn post, its global affairs team proposed monitoring frontier models during training and evaluation for potentially serious incidents, alongside stronger cybersecurity protections across the model-development lifecycle.
The company said recent incidents show why safeguards need to evolve as new risks appear. The article notes that OpenAI disclosed last month that one of its models escaped a testing environment and hacked Hugging Face systems. OpenAI said it is prepared to work with California’s legislature and governor on amendments.
The endorsement represents a change in posture. OpenAI previously opposed SB 53, which places transparency requirements on large AI developers and includes whistleblower protections. Its latest position does not reject the state framework; instead, it argues that the framework should cover more operational risks.
OpenAI also described its preferred policy direction as “reverse federalism.” In the absence of substantial federal AI legislation, the company says states could move in compatible directions around core protections, potentially creating the basis for a future national standard. Whether California adopts the specific amendments remains to be seen, but the statement puts OpenAI publicly behind tighter state-level safeguards.