The court files show efforts that have temporarily stopped licensing books to train artificial intelligence
New court files In the issue of publishing rights Amnesty International against Meta, add credibility to Previous reports The company “temporarily stops” with books publishers about licensing deals to provide some of the artificial intelligence models with training data.
Deposits are related to the case Kadrey V. Meta – One of these cases that belong to the American courts system that incited artificial intelligence companies against authors and other intellectual property holders. For the largest part, the defendants in these cases – artificial intelligence companies – claimed that training in copyright content is “fair use”. Prosecutors – copyright holders – did not agree noisy.
The new deposits submitted to the court on Friday, which include the partial texts of the deposits of the definition employees taken by the lawyers in the case, indicate that some identification employees felt negotiating the licenses of artificial intelligence training data for books that may not be developed.
According to one of the texts, SY Choudhury, who leads Meta partnership initiatives of artificial intelligence, said that Meta communicates with various publishers has met “very slow to link and interest.”
“I do not remember the entire list, but I remember that we have presented a long list of research at the beginning on the Internet of senior publishers, etc.” – from a lot of communication with cold calls to try to create the contact. “
“There were a few, like that, as you know, to engage, but not much,” Chaudri added.
According to court texts, Meta stopped some of AI’s books licensing efforts in early April 2023 after facing “timing” and other logistical setbacks. Chaudhry said that some publishers, especially the publishers of fictional books, did not actually have rights to the content that Mita is thinking about the license, according to the copy.
“I would like to point out that – in the fantasy category, we quickly learned from the business development team that most of the publishers we were talking to, they themselves representing that they had, in fact, the licensing rights and Chaudri said:” Data is us. “And so it would take it. A long time to interact with all their authors. “
Codari indicated that he was deposited that dead at least one occasion stopped licensing efforts to develop artificial intelligence, according to a copy.
“I am aware of the license efforts, for example, we tried to license three -dimensional worlds of various engines and engines of the artificial intelligence research “In the same way that I describe here for imagination and textbooks, we only received little participation in a conversation […] We decided – in this case, we decided to build our own solution. “
Prosecutors’ lawyer, among the best-selling authors, Sarah Silverman and Ta-Nehisi Coates, amended their complaints several times since the case was filed in the American District Court of the Northern District in California, the San Francisco section in 2023. Prosecutors claim that Meta is among crimes Other, some pirated books have been referred to with copyright books available for licensing to determine whether it is logical to follow a license agreement with a publisher.
The complaint also accuses dead Using “shadow libraries” that contain e -books pirate To train many artificial intelligence models of the company, including the famous Llama series of “open” models. According to the complaint, Meta may have got some libraries via torrent. Torring, a way to distribute files via the web, requires that the torrents at one time be “seeds”, or download, the files they are trying to get – Who confirmed the prosecutors is a form of copyright violation.
