Federal Judge Rules Anthropic Fair Use, Orders Piracy Trial

A federal judge in San Francisco ruled that Anthropic’s use of legally purchased and digitized books to train its AI chatbot Claude constitutes fair use under U.S. copyright law, marking a significant win for AI companies by recognizing the transformative nature of AI training. The judge compared the AI learning process to a student studying texts to create new work, emphasizing that the AI generates unique content rather than merely replicating copyrighted material. However, the court also found that Anthropic’s copying and storage of more than 7 million pirated books for its AI training library infringed authors’ copyrights, ordering a trial in December to determine damages. The lawsuit, filed by authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson, accuses Anthropic of large-scale theft of copyrighted works, though the company had shifted to purchasing copies after internal concerns about piracy. This ruling is the first to address fair use in generative AI, underscoring the legal complexities tech companies face in sourcing training data while affirming the transformative use defense. Backed by Amazon and Alphabet, Anthropic remains under scrutiny for its data acquisition methods despite the favorable ruling on AI training.