OpenAI is urging California to reinforce its AI safety framework by expanding key protections in SB 53, the state's landmark artificial intelligence bill.
In a statement from its global affairs team, the company said the law should go further by adding monitoring for frontier models during training and evaluation, especially to detect potential serious incidents early. OpenAI also recommended stronger cybersecurity protections across the full model-development process.
The company said it wants to work with California lawmakers and the governor to help shape a more resilient version of the bill. It also pointed to recent events as a reminder that safety rules must evolve as AI systems become more capable and more widely used.
The shift is notable because OpenAI had previously opposed SB 53, which includes transparency requirements and whistleblower protections for major AI companies. Now, the company says that in the absence of broad federal legislation, states can help build a compatible set of core protections that may eventually support a national standard.
This evolving approach suggests that AI governance may increasingly be shaped by state-level leadership, setting the stage for broader and more coordinated safeguards in the future.
