According to court documents, Meta has suspended efforts to negotiate licensing agreements with book publishers for its artificial intelligence (AI) models. This information is included in documents submitted to the court in connection with copyright infringement claims against Meta. Currently, Meta is facing several copyright claims. In one of the cases, authors claim that Meta illegally used their books to train its AI models. Meta, on the other hand, argues that the training process falls under the "fair use" principle.
According to the court documents, Meta's AI partnerships lead, Sy Choudhury, stated that the company tried reaching out to book publishers, but received little interest in return. While some publishers were open to collaboration, most could not offer licenses due to issues with copyright ownership.
Choudhury further noted that negotiations with fiction book publishers faced difficulties, as many publishers stated that they did not fully own the rights to the books. This meant separate agreements had to be made with each author, which is a time-consuming process.
Authors have accused Meta of using "shadow libraries," informal sources of electronic books, for training its AI models. They claim Meta used illegally obtained books and even downloaded them via torrent technology. Torrenting, which involves both downloading and sharing files, could be considered a violation of copyright.
The court is currently investigating whether Meta has infringed copyright laws. Authors claim that Meta compared pirated books with officially licensed ones to determine which materials to use. If these allegations are proven true, Meta could face serious legal and financial consequences.
