Scouting the battlefield: Will disputes over the interpretation of California's new law on AI data transparency render it meaningless? | Discussed By: Jason L. Haas
(February 04, 2025) California's new AI Training Data Transparency law, AB 2013, aims to shed light on the datasets used to train generative AI models. The law, which takes effect in 2026, requires developers to disclose key information about their training data, but its vague language is expected to spark legal battles. Companies that fine-tune AI models may also be subject to these disclosure rules, raising concerns about competitive secrecy and compliance. Courts will likely play a crucial role in interpreting AB 2013, particularly in defining what qualifies as a "high-level summary" of training data. While the law is intended to increase transparency and consumer confidence, developers may attempt to minimize disclosure through narrow readings of the statute’s requirements. Ultimately, how AB 2013 is enforced will determine whether it truly opens the AI "black box" or remains a largely symbolic measure.
As discussed by Jason L. Haas
Read more here Scouting the battlefield: Will disputes over the interpretation of California's new law on AI data transparency render it meaningless?
