A court has ordered Meta to identify the executives who were involved in torrenting books that were used to train its artificial intelligence models. The ruling, which compels the company to disclose names rather than shield them behind corporate anonymity, could shift legal accountability from the company itself to the individuals who oversaw the data acquisition.
The decision lands in the middle of a broader fight over how AI companies source the material they feed their models. Meta has previously acknowledged using copyrighted books in training, and the torrenting detail has been a flashpoint in litigation over whether that use was lawful.
Why the court wants names
The order doesn't settle whether Meta broke the law. It settles who has to answer for the conduct. By requiring the company to name the executives behind the torrenting, the court is treating the data-gathering process as something specific people decided to do — not as an amorphous corporate act. That distinction matters because lawsuits against AI firms have largely targeted the companies, leaving individual officers out of the frame. If executives can be named, they can be deposed, and potentially held personally responsible.
Meta has not publicly commented on the ruling. The company has consistently argued that its use of copyrighted material for training falls within accepted bounds, though it hasn't detailed the torrenting process in public filings.
What the ruling changes for corporate governance
For Meta's board and legal team, the immediate task is compliance: turning over the names by whatever deadline the court set. But the longer-term effect could be a rethinking of how AI data deals are documented and who signs off on them. Companies that have treated training-data acquisition as a low-level engineering problem may now need to treat it as a board-level risk. The ruling doesn't create new law, but it signals that courts are willing to look past the corporate entity when the conduct at issue is specific and traceable.
That has implications beyond Meta. Other AI developers that scraped or downloaded copyrighted works face similar exposure, and their executives may now think twice before approving data pipelines that rely on questionable sources.
The torrenting detail at the center of it
Torrenting is a peer-to-peer file-sharing method that has long been associated with piracy. Using it to assemble a training corpus is not the same as buying a license or striking a deal with publishers. The court's order suggests that the method — not just the fact of using the books — is what drew scrutiny. Meta will now have to explain who chose that method and who knew about it.
The company has not said whether it will appeal the order. If it does, the names could stay sealed for months while the appeal plays out. If it doesn't, the disclosure could come quickly.
What happens next
The court has set a schedule for Meta to comply with the order, though the exact date wasn't immediately clear from the ruling. Meta's legal team is expected to review the decision and decide whether to seek a stay. Meanwhile, the plaintiffs in the underlying case are likely to use any disclosed names to expand their discovery requests. The next hearing in the matter is expected to address the scope of the disclosure, and that's when the fight over how much Meta has to reveal will really begin.
