CSOAI - California AI laws
California AI laws - the patchwork
California has no single AI act; instead a set of targeted laws apply - training-data transparency, AI content disclosure, and automated-decision rules under privacy law. Together they form one of the toughest US environments.
What it covers
- !Generative-AI training-data transparency (AB 2013)
- !AI content disclosure / watermarking (SB 942)
- !Automated decision-making technology under CPRA
- !Sector rules (employment, health, elections)
What you must do
- +Publish training-data documentation for GenAI
- +Disclose and mark AI-generated content
- +Honor opt-outs and access rights for automated decisions
- +Run risk assessments under CPRA rulemaking
Questions, answered
California has no single AI act but several targeted laws, including AB 2013 (training-data transparency), SB 942 (AI content disclosure), and automated-decision-making rules under the CPRA.
AB 2013 requires developers of generative AI to publish documentation about the data used to train their systems.
Yes. California's privacy rulemaking covers automated decision-making technology, including access, opt-out, and risk-assessment expectations.
US state AI laws are evolving quickly. This reflects the position as of June 2026 and is not legal advice - verify current effective dates and requirements.