A PYMNTS Company

California’s SB 53: The First Frontier AI Law, Explained

By  |  October 13, 2025
California’s SB 53: The First Frontier AI Law, Explained

In this blog post, author Justine Gluck (Future of Privacy Forum) discusses California’s groundbreaking frontier AI legislation, SB 53, signed by Governor Newsom on September 29. The Transparency in Frontier Artificial Intelligence Act makes California the first state to specifically target frontier AI safety and transparency, requiring advanced AI developers to publish governance frameworks and transparency reports, establish safety incident reporting mechanisms, extend whistleblower protections, and develop a public computing cluster. While supporters view it as a critical step toward transparency and risk reduction, critics argue its requirements could burden developers and stifle innovation. The legislation comes as New York considers similar frontier AI legislation and serves as a potential blueprint for other states in the absence of comprehensive federal AI regulation.

    Get the Full Story

    Complete the form to unlock this article and enjoy unlimited free access to all PYMNTS content — no additional logins required.

    yesSubscribe to our daily newsletter, PYMNTS Today.

    By completing this form, you agree to receive marketing communications from PYMNTS and to the sharing of your information with our sponsor, if applicable, in accordance with our Privacy Policy and Terms and Conditions.

    The post explains that SB 53 regulates “frontier developers” who train high-compute AI models (those using more than 10^26 computational operations) and “large frontier developers” with revenues exceeding $500 million. The law focuses on preventing catastrophic risks, including mass casualties, assistance in creating weapons of mass destruction, autonomous criminal conduct, or models evading developer control. Key requirements include annual Frontier AI frameworks for large developers describing risk identification and mitigation, transparency reports before model deployment, mandatory reporting of critical safety incidents to the Office of Emergency Services within 15 days (or 24 hours for imminent dangers), and robust whistleblower protections. The Attorney General can enforce violations with penalties up to $1 million per incident.

    Gluck traces SB 53’s evolution from its predecessor SB 1047, which Governor Newsom vetoed last year. The revised legislation removed the most controversial provisions that critics deemed technically infeasible or barriers to innovation, including mandatory “kill switch” capabilities, pre-training safety requirements, annual third-party audits, strict 72-hour incident reporting windows, and steep penalties tied to compute costs. The streamlined SB 53 instead focuses on deployment-stage obligations with longer reporting timelines and capped penalties, representing a more balanced approach that addresses some but not all criticisms while establishing California’s leadership in frontier AI regulation…

    CONTINUE READING…