California enacts AI Transparency Act, setting new rules for generative AI

California has established new rules for artificial intelligence, enacting the AI Transparency Act (SB 942) to govern generative AI systems that reach a substantial number of users. Governor Gavin Newsom signed the bill into law in 2024, with its provisions officially taking effect on August 2, 2026.

The legislation, also known as the California AI Transparency Act, adds a new chapter to the state's Business and Professions Code. This chapter, starting at section 22757, aims to increase transparency in generative AI systems. The law defines artificial intelligence as machine-based systems capable of autonomy that generate outputs influencing physical or virtual environments based on received information. Generative AI is specifically defined as systems producing synthetic content, such as text, images, videos, and audio, that mimic the training data.

Initially set to take effect on January 1, 2026, the law's implementation was postponed to August 2, 2026, through a subsequent amendment, AB 853. This complementary law, signed by Newsom on October 13, 2025, broadened the scope of the original AI Transparency Act. It introduced additional requirements for large digital platforms, distributors of generative AI systems, and manufacturers of devices that capture photos, audio, and video.

AB 853 expanded the definitions within the legislation to include categories not originally covered by SB 942. A "capture device" is now defined as any equipment that can record photos, audio, or video, encompassing cameras, mobile phones, and voice recorders. The law also introduced the category of "generative artificial intelligence system hosting platform," referring to sites that distribute such systems.