"Meta Platforms Inc. faces a new proposed class action from several major book publishers and authors who say its Llama generative AI was trained through “one of the most massive infringements of copyrighted materials in history.”

Companies including Elsevier Inc., Cengage Learning Inc., and Hachette Book Group Inc. filed the latest in content owner suits over AI training Tuesday in the US District Court for the Southern District of New York.

The complaint alleges Meta made unauthorized copies of millions of books and articles through torrenting, web-scraping, and training—a peer-to-peer file-sharing method—while also illegally removing copyright information from them.

The filing, which also names CEO Mark Zuckerberg as a defendant, attacks Meta’s use of the works at issue at different stages based on a number of theories, reflecting the legal uncertainties surrounding the application of copyright law to generative AI.

It addresses initially torrenting the works from illicit “pirate libraries” like Anna’s Archive, LibGen, and Sci-Hub, an intermediary stage where copies are made during the training process, and the end-game in which Meta is “flooding the market with AI-generated substitutes” for their works...."

This article was originally published in Bloomberg Law.