Fact Check Team: Anthropic court rulings could shape AI copyright future

Anthropic faces a new lawsuit from over 100 authors, alleging the use of pirated books to train its Claude AI system.
The authors claim Anthropic used unlawfully obtained digital copies for its training dataset, contributing to a wave of litigation against AI companies.
A key legal question at issue is whether AI companies can legally train on copyrighted material and if the method of obtaining such material impacts this.
In a federal case in California, a judge delineated that training an AI model on legally acquired books may qualify as “fair use” but did not address the usage of pirated material.
The fair-use ruling left unresolved issues regarding allegations of downloading and storing pirated books, indicating ongoing disputes over copyright liability.
Courts are now breaking down these disputes into narrower legal questions rather than making broad rulings on generative AI as a whole.
The case’s resolution will depend on a separate trial focusing on the piracy-related claims.
Similar lawsuits against multiple AI developers and cases involving music publishers are examining whether fair-use principles apply broadly to creative works beyond just books.