Major Publishing Houses Sue Book Piracy Site WeLib

A group of major publishers filed a lawsuit against “endless library” and alleged piracy website WeLib on Tuesday, claiming that the site engaged in rampant copyright infringement and allowed artificial intelligence developers to illegally use its content to train large language models. 

The group comprises 13 members of the Association of American Publishers, including the “Big Five” publishing houses: Penguin Random House, Macmillan, Simon & Schuster, Hachette Book Group, and HarperCollins. Other members include academic publishers Cengage, McGraw Hill, and Wiley.

“Publishers bring this action to stop the mass distribution by Defendants [WeLib] of millions of legally protected literary works owned by Publishers that were unlawfully copied from physical and digital books and journals,” the lawsuit states. “Publishers’ action is now especially critical in light of reports that major large language model developers used illegal sites like WeLib as illicit sources of training data.”

WeLib is an online repository of books, journal articles, and other materials comprising at least 141 million documents, according to its website. 

The plaintiffs’ investigators searched WeLib’s database for the name of each publishing house and found hundreds of thousands of search results. For example, a search for “Penguin Random House” yielded 639,586 results available for download; another search for “McGraw Hill” yielded 137,138 results. The investigators were also able to download full copies of 130 specific works owned by the publishers from the website. 

“The scale of WeLib’s infringement is stunning,” the publishers wrote. WeLib did not immediately respond to a request for comment. 

This lawsuit comes on the heels of another legal challenge last month, when the same 13 publishers won a lawsuit against piracy website Anna’s Archive for the same allegations as in this case. Anna’s Archive was ordered to be shut down. Publishers have also recently filed a class action lawsuit against Meta for “willful copyright infringement” to develop its Llama AI models.