OpenAI is urging California lawmakers to expand SB 53 with stronger frontier-model monitoring and cybersecurity rules, marking a shift from its opposition before the law was signed in 2025.
Key Points:
- OpenAI wants frontier models monitored during training and evaluation for serious security-related behavior.
- The company also supports stronger cybersecurity requirements across the model-development lifecycle.
- California's SB 53 already requires safety disclosures, incident reporting and whistleblower protections.
OpenAI SB 53
OpenAI's Global Affairs team said California should amend the Transparency in Frontier Artificial Intelligence Act to require monitoring for conduct that could bypass third-party security controls or compromise confidential information.
The company also called for stronger cybersecurity protections throughout model development, particularly safeguards designed to stop frontier models from circumventing internal security controls. OpenAI said the goal is to detect dangerous behavior earlier, respond faster and share lessons that could improve safety across the industry.
The shift is significant. OpenAI lobbied against SB 53 before Gov. Gavin Newsom signed the bill on Sept. 29, 2025, but the company now describes the law as an important foundation for frontier AI safety.
The existing framework is already broad. SB 53 requires large frontier developers to publish safety frameworks and report critical safety incidents, while protecting whistleblowers who disclose major health or safety risks. The law also gives California's attorney general authority to seek civil penalties for violations and created the CalCompute public computing initiative.
Also Read: Pi Network Holds Bullish Triangle, But $0.085 Breakdown Risk Grows
California AI Policy
OpenAI has linked its support for stronger state rules to a strategy it calls “reverse federalism,” under which states establish compatible safeguards that could later help shape a national framework.
Chris Lehane, OpenAI's chief global affairs officer, wrote in July that California, New York and Illinois had begun moving toward a shared baseline for governing the most powerful AI systems. OpenAI argues that aligned state laws could give federal policymakers a foundation for national standards without freezing safety requirements as technology changes. That approach puts states at the front of AI rulemaking.
California designed SB 53 to evolve with the sector. The law directs the California Department of Technology to recommend updates each year based on technological developments, international standards and stakeholder input, giving lawmakers a formal mechanism to revisit frontier-model safeguards as new risks emerge.
Read Next: Term Finance Loses $8.5M In Ethereum Governance Attack
