Meta Achieves Mixed Fair Use Win in AI Copyright Dispute * TorrentFreak

In recent years, rightsholders have increasingly sued AI developers over alleged copyright infringements during the training of language models. Meta, facing a class-action lawsuit from authors like Richard Kadrey and Sarah Silverman, admitted using pirated books for training its Llama model but claimed fair use. A recent court ruling favored Meta, denying the authors’ motion to hold them liable for copyright infringement. The court emphasized that no significant market harm was caused, dismissing claims that Llama could replicate substantial portions of the authors’ work or harm a nascent AI licensing market. However, it recognized potential challenges in future cases regarding market dilution, where AI-generated content might compete with original works. The ruling, while a win for Meta, underscores that it doesn’t set a legal precedent for future cases and suggests that proper licensing might soon be necessary for AI training practices. Judge Chhabria noted that the authors’ arguments were inadequately supported.

Source link