Assembly Member Bauer-Kahan's legislation addressing copyrighted materials in artificial intelligence training would establish new documentation and disclosure requirements for developers of generative AI systems in California. The bill creates a framework requiring developers to track and report the use of copyrighted content in their AI training data.

Under the proposed requirements, developers must document all copyrighted materials used to train their generative AI systems before making them available to California users. The documentation must identify copyright owners and be retained for the duration of the system's commercial use plus 10 years. Developers must also create a mechanism on their websites for copyright owners to submit inquiries about the use of their materials.

The bill sets specific timelines for developer responses to copyright owner requests. When copyright owners whose materials were used in training submit written requests, developers must provide comprehensive lists of those materials within seven days. For inquiries from copyright owners whose materials were not used, developers must confirm this within 30 days. Copyright owners who do not receive timely responses may pursue civil actions seeking damages of $1,000 per violation or actual damages, whichever is greater, along with injunctive relief and attorney's fees.