Zuckerberg gave Meta Llama team OK to train on copyrighted works, filing claims

Plaintiffs in the copyright lawsuit Kadrey v. Meta allege, in newly unredacted court filings, that CEO Mark Zuckerberg approved the Llama AI team's use of LibGen, a dataset of pirated books and articles, for training. According to Meta's testimony as relayed by plaintiffs' counsel, the approval came despite internal concerns, with one employee describing LibGen as a "data set we know to be pirated." The filings also allege that Meta stripped copyright information from training data to conceal infringement and torrented LibGen, which plaintiffs' counsel calls another form of infringement. Meta has not yet responded, and the case remains undecided; Meta argues fair use.