California delays SB 942 to August, expands AI transparency law

On January 7, California lawmakers passed AB 853. The bill delays SB 942 until August 2, 2026. It also broadens disclosure rules for entities that host or deploy generative AI and for large online platforms. Reports from IAPP and The Verge describe the package as postponing enforcement timelines and widening the set of covered actors. (IAPP, Jan 7; The Verge, Jan 1)

The bill adds or strengthens disclosure duties for AI hosting platforms and large online services. It increases public and regulator visibility into model deployment and safety practices. Coverage from the Santa Barbara Independent and IAPP notes these transparency measures are part of a wider 2025 legislative sweep covering AI and privacy. (IAPP, Jan 7; Independent, Jan 7)

What’s next: Platforms that host generative models have a multi-month window to prepare before the rules take effect. Treat the August 2, 2026 date as firm for planning. Legal, product, and security teams should inventory models, prepare disclosures and risk documentation, and document mitigation and incident response plans now to avoid rushed remediation.

Why It Matters

• Inventory models, training data sources, and safety testing now—August 2, 2026 is the compliance deadline.
• Legal and privacy teams should prepare disclosure templates and public notices to meet broader transparency obligations.
• Security and ML teams must document mitigation measures and incident response plans to reduce exposure during regulatory review.