The plaintiffs—Elsevier, Cengage, Hachette Book Group, Macmillan Publishers, McGraw Hill, and Scott Turow—claim Meta obtained material from pirate sites and stripped copyright management information to conceal origins before using it for AI training.

The outcome of this case will influence AI development practices across the industry and could drive licensing agreements or establish new precedents on fair use and AI training.

There remains broad uncertainty about fair use in AI training, as courts have not issued a definitive ruling on the issue.

Publishers and author Scott Turow filed a copyright infringement lawsuit against Meta on May 5, 2026, accusing Meta of illegally downloading millions of books and journal articles to train its Llama AI models.

The complaint alleges Meta torrented over 267 terabytes of copyrighted material from piracy platforms, including textbooks, research papers, novels, and professional publications used without permission or payment.

The central legal question is whether AI training on copyrighted material counts as fair use or requires licensing, with Meta likely arguing that much of the material was publicly available online and that AI training can be fair use.

The case could set major precedents for the AI industry: a Meta win could enable continued training on publicly available data without compensation, while a publishers’ win could raise licensing costs for training models.