Authors Sue Anthropic for Training AI Using Pirated Books

A group of authors have sued Anthropic, accusing it of training its models based on pirated books, as reported Reuters Agency. This proposed class action lawsuit The lawsuit was filed in a California court on Monday and accuses Anthropic of “building a multibillion-dollar business by stealing hundreds of thousands of copyrighted books.”

“It is clear that Anthropic downloaded and reproduced copies of The Pile and Books3 knowing that these datasets consisted of a vast amount of copyrighted content from pirated websites such as Bibiliotik,” the lawsuit reads. The authors want the court to certify their class action lawsuit, as well as demand that Anthropic pay proposed damages and bar the company from using the copyrighted material in the future. Anthropic did not immediately respond to the complaint. EdgeRequest for comment.

Among the authors suing Anthropic is Andrea Bartz, author We were never here; Charles Graeber, who wrote Good Nurse; and Kirk Wallace Johnson, author Feather Thief. While the lawsuit confirms that Books3 was removed from the “most official” version of The Pile, the original version is allegedly still available elsewhere online. A recent investigation also found that companies like Anthropic and Apple trained their AI models on thousands of scraped YouTube video captions available on The Pile.