Meta Ordered to Disclose Who Approved Pirated Downloads for AI Training
Understand the story
In the case of authors suing Meta for training Llama models on pirated books, the court in October 2026 ordered Meta to disclose the executives who approved torrented downloads of pirated works for the early AI training of models such as Llama 1 and Llama 2, along with their roles and timing of involvement. Meta argued that leadership made no formal decisions at the time and that the approvals came from collective discussions. Judge Hixon rejected this claim, noting that the plaintiffs were asking about all decision-makers, not just formal approvers. Additionally, Meta was required to explain what happened to the pirated works that were downloaded but not used for training. The case has now reached the court-ordered disclosure stage, with Meta required to submit the relevant executive information and explanations.
AI overview of reporting · Updated 2 h前
Reporting timeline
Explore different aspects of the story through its reporting.
- Ars Technica · AIMeta tries to scapegoat staff who downloaded torrents to spare execs
A court has ordered Meta to disclose the executives who approved using pirated works downloaded via torrents for early AI training of models like Llama 1 and 2, and to explain their roles and timing of involvement. Meta argued that leadership made no formal decisions and approvals came from collective discussions; Judge Hixon rejected these claims, noting the authors are asking about all decision-makers, and Meta must also account for pirated works not used in training.
Story heat trend
Not enough continuous observations to show a trend yet.