A federal magistrate judge has ordered Meta to identify the supervisors, managers, and executives who authorized the downloading of pirated books from shadow libraries like LibGen and Anna’s Archive. The decision stems from copyright class-action lawsuits brought by authors and publisher Entrepreneur Media. The order follows testimony from Meta employees indicating that executive approvals were required to download large dataset torrents used in training the company’s AI models.

Meta previously claimed an “informal approach” to dataset acquisition, arguing in court filings that no central authority made explicit decisions to pirate material, framing the action as technical downloads by individual staffers. However, internal communications unsealed during discovery revealed employee discomfort with corporate torrenting and indicated that dataset downloads were escalated to top leadership, including CEO Mark Zuckerberg.

In addition to naming decision-makers, the ruling forces Meta to produce server command history files from virtual machines and AWS instances. These log files could provide plaintiffs with concrete evidence of whether Meta seeded copyrighted works, supporting claims of contributory infringement.

Why it matters

  • Increases legal exposure for AI executives personally tied to unauthorized dataset acquisition decisions.

  • Demonstrates courts’ willingness to force disclosure of technical server logs to prove copyright infringement in AI training.

  • Signals rising compliance risks for startups relying on unvetted or shadow library training data.

Source: arstechnica.com