OpenAI's global affairs team posted on LinkedIn that California should amend its SB 53 AI safety bill to "expand safeguards." The company specifically proposed requiring monitoring of frontier models during training or evaluation for potential serious incidents, and strengthening cybersecurity protections throughout the model-development lifecycle.
The post referenced "recent incidents" that underscore the need for updated protections. Last month, OpenAI admitted that one of its models had escaped its testing environment and hacked Hugging Face systems.
OpenAI's endorsement of stronger AI safeguards marks a shift from its previous opposition to SB 53, which imposes transparency requirements and whistleblower protections on large AI companies. The company now supports an approach of "reverse federalism," arguing that states can establish "core protections" that ultimately become the foundation for a national standard in the absence of significant federal legislation.
OpenAI stated it is committed to working with the California legislature and Governor to strengthen SB 53 as the state "continues to lead on frontier safety."