Court orders Meta to name executives who authorized AI training data piracy
A magistrate judge has ruled that Meta must identify the leadership officials who approved the illegal downloading of copyrighted books for artificial intelligence training.
Key points
- Magistrate Judge Thomas Hixon ordered Meta to name key decision-makers involved in authorizing the piracy of books from shadow libraries.
- Meta previously argued that no executive made a formal decision, claiming staff merely downloaded torrents through an informal process.
- Internal messages revealed that some employees expressed discomfort with the directives and noted that higher-level approval was required.
Meta has been ordered by a court to identify the specific executives who authorized the company’s employees to download copyrighted books from shadow libraries for artificial intelligence training. This ruling, reported by TorrentFreak, comes after Magistrate Judge Thomas Hixon largely agreed with the plaintiffs that Meta must clearly name the key decision-makers behind these actions. The order impacts a class-action lawsuit filed by book authors, including Richard Kadrey and Sarah Silverman, as well as a separate lawsuit from publisher Entrepreneur Media.
The judge’s decision follows testimony from Meta staff members who admitted to downloading the torrents. These employees indicated that supervisors, managers, and executives had greenlit the piracy efforts. According to the court filing, the discovery order could expose a wider range of leadership personnel than Meta had initially attempted to shield. The ruling challenges Meta’s previous defense strategy, which sought to avoid detailing the chain of command involved in the early stages of its AI training data compilation.
Meta’s Defense Strategy
Meta had argued in court filings that its initial efforts to compile massive AI datasets were managed through an informal approach where no one with authority made any formal decisions. The company insisted that a decision was only made in the technical sense that some staffers downloaded the torrents. This argument aimed to shift blame away from leadership and onto individual employees who performed the downloads. However, the court rejected this characterization, requiring Meta to provide clear identification of those who authorized the activity.
Authors and publishers have contended that Meta made a willful choice to induce infringement by downloading and illegally distributing their copyrighted works. They argue that the scale of the downloads—80.6 terabytes from LibGen and 35.7 terabytes from Anna’s Archive—could not have been achieved quickly without the company also seeding the files. Meta has countered that it relied on mitigations to prevent seeding, but the plaintiffs maintain that the technical reality of torrenting supports their claim of intentional infringement.
Internal Pushback
Internal communications revealed during the lawsuits show that some Meta staff expressed discomfort with the directives to pirate copyrighted material. One worker commented that torrenting from a corporate laptop did not feel right. Another staffer testified that he initiated downloads that required director or vice president-level approval to be used for training. A joint letter from Entrepreneur Media and Meta highlighted this testimony. Additionally, another employee testified that leadership had previously halted and then resumed downloading torrents sourced from Anna’s Archive, suggesting active management oversight.
What should you do?
Security and technology professionals should monitor legal developments regarding corporate liability for AI training data sourcing. Organizations developing AI models should review their data acquisition policies to ensure compliance with copyright laws. Legal teams should assess whether current data collection methods involve authorized use of copyrighted materials. Companies should document decision-making processes for data sourcing to clarify responsibility and compliance.
What to do and how to stay safe: Meta
- Review internal policies regarding the sourcing of training data for artificial intelligence models to ensure legal compliance.
- Document all decisions related to data acquisition, including approvals from management, to establish a clear chain of command.
- Monitor legal proceedings involving AI training data to understand emerging liability standards for corporate data practices.
- Consult legal counsel to assess whether current data collection methods involve unauthorized use of copyrighted materials.
Step-by-step guide: Prompt Injection Warning Signs: How to Spot AI Manipulation
General security guidance from the Threat Times newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
What did the court order Meta to do?
The court ordered Meta to name the executives and key decision-makers who authorized the downloading of copyrighted books for AI training.
How did Meta defend its actions in court?
Meta argued that no executive made a formal decision, claiming staff downloaded torrents through an informal process without authoritative direction.
What evidence contradicted Meta's defense?
Internal messages and employee testimony showed that staff required higher-level approval for downloads and expressed discomfort with the directives.
