The case highlights the tension between using copyrighted material for AI training and the concept of "fair use," which remains unresolved in American courts. Despite Meta's attempt to keep these details private, Judge Vince Chhabria denied the request, noting that the company aimed to avoid negative publicity rather than protect sensitive business information. The filings reveal internal discomfort among employees about using pirated content, indicating awareness of the potential legal and reputational risks.
Key takeaways:
- Meta is accused of using pirated materials to train its Llama AI, with internal documents suggesting Mark Zuckerberg approved the use of a pirated dataset called "LibGen".
- The case, Kadrey v. Meta, involves authors like Sarah Silverman and Ta-Nehisi Coates, who claim their works were used without authorization.
- Meta employees expressed concerns about the legality and ethics of torrenting pirated materials on corporate devices.
- The legal argument hinges on whether the use of these materials falls under "fair use" protections, a question yet to be fully resolved by American courts.