Zuckerberg Approved Meta's Use of Pirated Data to Train Llama, Lawsuit Claims

Meta CEO Mark Zuckerberg allegedly approved the use of a dataset called LibGen, containing pirated ebooks and articles, to train Meta's Llama AI models. Plaintiffs, including bestselling authors Sarah Silverman and Ta-Nehisi Coates, accuse Meta of copyright infringement. Meta claims fair use, but plaintiffs argue Meta attempted to conceal infringement by stripping copyright information and even using torrenting to obtain LibGen, exacerbating the violation. The case currently only pertains to Meta's earliest Llama models, and the outcome remains uncertain, but the allegations have already damaged Meta's reputation.