New court deposits In a copyright affair AI against Meta add faith to previous reports That the company “ceases” with book publishers on licensing agreements to provide some of the AI genetic models with training data.
Deposits are related to the case Kadrey v. Meta Platforms – One of the many such cases that evolve through the US judicial system installed AI companies against writers and other intellectual ownership holders. For the most part, the defendants in these cases – AI companies – claimed that the training in copyright -protected content is “fair use”. The plaintiffs – owners of copyright – disagree with voice.
New deposits submitted to court on Friday, which include some transcripts of META employees’ deposits obtained by lawyers for the plaintiffs in the case, suggest that some post -personal negotiations have been negotiated for the training of AI Training Data for books for books for books for books. not to be scaled.
According to one copy, Sy Choudhury, who is leading META’s partnership initiatives, said Meta’s promotion to various publishers met with “very slow recruitment of engagement and interest”.
“I don’t remember the whole list, but I remember we had a long list of the first cleaning of the top publishers, et cetera,” Choudhury said, according to the copy, “and we did not get in touch and comments from – from many from our cold calls to try to create contact.
Choudhury added: “There were some, like this, you know, to get involved, but not many.”
According to copies of the court, Meta stopped some of the licensing of books related to AE in early April 2023 after the “timetable” and other logistical support failures were tackled. Choudhury said that some publishers, in particular, publishers of fiction books, have proven to have not actually had the rights to the content that Meta examined the licensing, per transcription.
“I would like to point out that the category of fiction, we have quickly learned from the business development team that most of the publishers we were talking about, they represented that they did not actually had the rights to use the data for us,” Choudhury said. “And so it would take a long time to deal with all their writers.”
Choudhury noted during his testimony that Meta has at least another opportunity to stop the licensing efforts associated with the development of AI, according to one copy.
“I am aware of the licensing efforts, such as, for example, we have tried to grant 3D Worlds license from different toy machine and toy manufacturers for our AI research team,” Choudhury said. “And in the same way I describe here for fiction and books. We have very little commitment to have a conversation […] We decided to – in this case, we decided to build our own solution. ”
The plaintiff’s adviser, including the authors of Sarah Silverman and Ta-Nehisi Coates, have amended their complaint several times since the case was filed with the US District Court for the Northern District of California, the San Francisco section. in 2023. Submitted by the plaintiff’s adviser claims that Meta, among other offenses, interrupted some pirate books with copyright -protected books to determine if it was reasonable to pursue an authorization agreement with a publisher.
The complaint also accuses META of using “shadow libraries” containing pirate e -books to train many of the company’s AI models, including the popular Llama series of “open” models. According to the complaint, Meta may have secured some of the libraries through torrenting. Torrenting, a way of distributing files across the web, requires torrenters at the same time “seeds” or upload, the files they try to obtain – which the plaintiffs claim is a form of copyright violations.
