The lawsuit highlights the ongoing legal challenges surrounding the use of copyrighted content in training AI models, a contentious issue in the development of generative AI tools. A US district judge previously dismissed claims that Meta's AI-generated text infringed copyrights and unlawfully stripped copyright management information (CMI) but allowed the plaintiffs to amend their claims. The authors have now presented new evidence to support their infringement claims and seek to revive their CMI case while adding a computer fraud allegation. The judge has permitted the filing of an amended complaint but expressed doubts about the fraud and CMI claims.
Key takeaways:
- Mark Zuckerberg allegedly approved Meta's use of pirated books from the LibGen dataset to train AI models, according to a court filing by a group of authors.
- The authors, including Ta-Nehisi Coates and Sarah Silverman, are suing Meta for copyright infringement, claiming their books were misused to train Meta's Llama language model.
- The use of copyrighted content in AI training is a contentious legal issue, with concerns about the impact on creative professionals and publishers.
- A US district judge allowed the authors to amend their claims, though expressed skepticism about the merits of their fraud and copyright management information allegations.