Required actions
- Post on the developer's website a high-level summary of the datasets used to train the generative AI system: sources, whether they include personal or copyrighted data, licensing, collection dates and whether synthetic data was used
Requires developers of generative AI systems available in California to publish a summary of their training data, including sources, personal data and copyrighted material, from 1 January 2026. A model-level transparency duty with broad reach.
Spot an error, or know a record we should add? Members can propose additions and submit corrections.