Meta Faces Backlash for Using Pirated Books in AI Training
December 2023 archive — Authors challenged Meta Platforms over the use of their books to train its Llama artificial-intelligence models. The dispute raised questions about copyright, the sources of training material, and the relationship between technology companies and writers.
In Kadrey v. Meta Platforms, authors alleged that Meta copied their books without permission. A later court order dated June 25, 2025 describes Meta’s use of online shadow libraries and the copyright arguments made by both sides. Allegations of infringement and a court’s findings about fair use are different matters.
The use of copyrighted books for AI training is a complex legal issue. Authors and publishers have expressed concerns over the unauthorized use of their works, which they view as a violation of their rights and a potential threat to their livelihoods. The incident has prompted discussions about the need for clearer regulations and guidelines in the rapidly evolving field of AI technology, especially concerning the use of copyrighted content.
The dispute also affected the relationship between technology companies and creative industries. Authors’ concerns about permission, credit, and compensation sit alongside developers’ arguments about how training systems use existing works.
The situation is a stark reminder of the challenges posed by emerging technologies and their intersection with established legal frameworks. As AI continues to advance, the need for a balanced approach that respects both innovation and the rights of content creators becomes increasingly crucial.
Archive update — October 8, 2026: On June 25, 2025, the court granted Meta summary judgment on the training-related copyright claim brought by the thirteen authors in that case. The judge stressed that the decision depended on the evidence presented and did not establish that all such AI training is lawful. Read the court’s order. This dated note replaces the original statement that Meta had not responded; it is not a complete account of subsequent litigation.
Any content used for AI needs an attribution (A series of tags and links demonstrating the source and license for a digital material product’s use by the creator or the copyright holder) Until an attribution system is established including a public domain register for each country, we will never be certain that the materials used are legal, legitimate, and copyright free or have had their copyright use authorized and paid for. Every item used in an AI training model must be attributed.