Meta is facing a class-action lawsuit filed by five major publishers and an author, alleging that the company “engaged in one of the largest copyright violations in history” while training its AI Llama models, as The New York Times. In their lawsuit, Macmillan, McGraw-Hill, Elsevier, Hachette, Cengage , and author Scott Turow allege that Meta “repeatedly copied” their books and scholarly articles without permission.
See also: Meta introduces AI bone structure analysis to identify minors
The lawsuit accuses Meta of knowingly copying copyrighted works from “notorious pirate sites,” such as LibGen, Anna’s Archive, Sci-Hub, Sci-Mag , and others, and then feeding that material into its AI model. It also alleges that Meta trained Llama with information from the Common Crawl dataset, which is allegedly “filled with unauthorized copies of copyrighted works.”
As a result, Llama “produces verbatim and near-verbatim substitutes” for copyrighted material. For example, when asked for two short sentences from Cengage’s best-selling textbook, Calculus: Early Transcendentals, 9th edition, by James Stewart, Llama begins to reproduce the rest of the section verbatim.
Several authors have already sued Meta for alleged copyright infringement, which has brought to light the company's internal discussions about how to handle "media coverage suggesting that we have used a dataset that we know is pirated."
See also: Meta loses historic child safety trial
Last year, a federal judge ruled in Meta's favor in one of these lawsuits, though he noted that his decision "does not mean that Meta's use of copyrighted materials to train its language models is legal."
A group of authors also sued Anthropic for copyright infringement. While a federal judge ruled that training AI models on legally purchased books without a license constitutes fair use, he allowed the authors to move forward with a class action lawsuit over the “millions” of works that Anthropic allegedly used. Anthropic agreed to pay the authors $1.5 billion last year to settle the class action.
Turow and the group of publishers are suing Meta for damages and asking the court to order the company to stop its allegedly illegal activities. They are also asking the court to require the company to provide a list of books, scientific articles and other copyrighted works on which it trained its AI Llama models.
See also: Meta: Impressive growth of business AI
“AI fuels transformative innovation, productivity, and creativity for individuals and companies, and courts have rightly found that training AI on copyrighted material can be considered fair use,” said Meta spokesman Dave Arnold.
