Meta Ordered to Name Executives Who Approved Torrenting for AI Training
Meta faces a court order to identify specific executives who approved the use of pirated books for AI training, while the company has argued that nobody officially 'decided' to perform the torrent downloads. This legal development highlights the growing corporate accountability issues surrounding copyright infringement in AI data sourcing and challenges efforts by upper management to deflect blame onto lower-level employees. Court proceedings revealed that Meta attempted to scapegoat staff members who physically downloaded the torrent files, claiming no single executive made a deliberate choice to authorize copyright infringement.
## BACKGROUND
AI companies frequently scrape massive datasets, including books and internet articles, to train large language models, often leading to copyright lawsuits from authors and publishers. Torrenting involves peer-to-peer file sharing networks frequently used to distribute unauthorized copyrighted media.