United States of America: Act on frontier artificial intelligence models and whistleblower protections including transparency requirements (SB 53) was adopted by California Legislature

On 13 September 2025, the California Legislature adopted the Transparency in Frontier Artificial Intelligence Act (SB 53). The Law applies to frontier developers and large frontier developers of foundation models, including those with annual revenue above USD 500 million and models trained with computing power exceeding a certain threshold. The Law requires large frontier developers to publish and annually update a frontier AI framework, issue transparency reports before deploying new or substantially modified frontier models, and transmit summaries of internal catastrophic-risk assessments to the Office of Emergency Services, with critical safety incidents reported within 15 days. Furthermore, the Law creates whistleblower protections for covered employees, requires an anonymous internal reporting channel with periodic status updates, and authorises civil penalties up to USD 1 million per violation, while pre-empting local rules on catastrophic-risk management and exempting specified reports from public records disclosure. The Law also establishes a consortium within the Government Operations Agency to design “CalCompute,” a public cloud computing cluster, with a framework report due to the Legislature by 1 January 2027 and operation subject to appropriation.