Facebook employees used Torrent to download 81.7TB of data to train its LLaMA: Court papers

Meta, Facebook’s parent company, faces a class action lawsuit alleging copyright infringement and unfair competition in training its LLaMA AI model, according to court records revealed by vx-underground on X. First reported by Tom’s Hardware, vx-underground’s post on Twitter claims how court documents from February 5th, 2024, in Kadrey v. Meta show Meta (formerly Facebook) illegally Torrented 81.7TB of data from “shadow libraries” such as Anna’s Archive, Z-Library, and LibGen to train Meta artificial intelligence.

The lawsuit claims that internal communications show Meta researchers expressing concerns about using pirated material as early as October 2022, with one senior AI researcher stating, “I don’t think we should use pirated material.