Meta must identify people who made, authorized, or approved decisions to obtain disputed book datasets from shadow libraries, according to a discovery order in copyright litigation brought by authors and a publisher.
The cases concern downloads from sources including LibGen and Anna’s Archive for AI training. Plaintiffs argue that Meta could not have downloaded 80.6 terabytes from LibGen and 35.7 terabytes through Anna’s Archive without also distributing material through BitTorrent. Meta says it used measures intended to prevent seeding.
Magistrate Judge Thomas Hixon largely accepted requests to identify decision-makers and ordered Meta to produce command-history files for servers, virtual machines, and AWS instances used for torrenting. Those logs could show how the downloads operated and whether files were uploaded to other peers.
Meta has described the early data-gathering process as informal and disputed the scope and wording of the plaintiffs’ requests. The order is a discovery ruling, not a finding that Meta infringed copyright or that any named executive is liable. It may, however, clarify who approved the acquisition of training data and how those instructions were carried out.