A new class action filed by Lieff Cabraser Heimann & Bernstein and Susman Godfrey on behalf of several authors alleges that artificial intelligence startup Anthropic committed “brazen infringement” by using “hundreds of thousands” of copyrighted books to train its flagship collection of large language models, “Claude.”
The complaint, filed on Monday in the Northern District of California, accuses the San Francisco-based company of illegally downloading and copying pirated versions of the material to feed to its LLMs, which are designed to simulate human communication and generate predictive written responses to prompts by algorithmically processing the datasets they ingest.
