Bill > AB412

This bill establishes new requirements for developers of generative artificial intelligence (GenAI) models regarding the use of copyrighted materials in their training data. The legislation requires developers to document any copyrighted materials they knowingly use in training their AI models and create a mechanism on their website that allows rights owners (copyright holders) to request information about potential use of their copyrighted content. Specifically, developers must provide a way for rights owners to submit fingerprints or identification numbers of their copyrighted works, and within 30 days, the developer must assess whether those materials are likely present in their training dataset and provide a list of identified materials. If a developer fails to respond or comply, rights owners can bring a civil action and potentially receive up to $1,000 per violation, along with potential injunctive relief and attorney's fees. The bill does not apply to AI models trained exclusively on publicly available data, used for academic or governmental research, or developed without using copyrighted materials. The legislation aims to increase transparency and provide copyright holders with a mechanism to understand and potentially challenge the use of their intellectual property in AI training datasets, reflecting growing concerns about copyright issues in artificial intelligence development.

Business and Industry, Justice

https://leginfo.legislature.ca.gov/faces/billStatusClient.xhtml?bill_id=202520260AB412