Authors ask judge to focus on Meta torrent uploads in Llama case

In short: Authors suing Meta want to update their lawsuit to highlight that Meta allegedly shared pirated books through BitTorrent while collecting data to train its Llama AI model.

What happened

A group of authors in a class-action copyright case against Meta asked the court for permission to amend, or rewrite, their complaint. They want to emphasize Meta’s alleged use of BitTorrent to get copies of their books from “shadow libraries” such as Anna’s Archive.

BitTorrent is a file-sharing system where downloading and uploading usually happen at the same time (like getting a stack of papers while also handing out copies to other people in the room). The authors say this matters because it could mean Meta distributed copies of their books to other BitTorrent users, not just downloaded them.

Meta recently filed a response arguing that any uploading during BitTorrent downloads should be treated as fair use. Fair use is a legal rule that can allow limited copying in certain situations. Meta says the uploading is built into how BitTorrent works and was part of an efficient process to obtain a large dataset for AI training.

This dispute comes after the judge previously found that Meta’s training itself, even using pirated books, qualified as fair use. That leaves the alleged BitTorrent sharing and distribution as a main remaining issue in the case. Judge Vince Chhabria has not yet ruled on whether Meta can rely on this fair use argument for the uploading.

Why it matters

This case could influence how courts treat AI companies that collect training data through torrent-style sharing, especially if the process also sends files to other people. It may also affect other lawsuits involving shadow libraries and torrenting.

Source: Arstechnica