Meta Probes Employees Tied to Adult Film Downloads
Meta says its internal search found staff who appear to have interacted with Strike 3 videos, a claim it says is unrelated to its AI training.
Updated on Nov. 5, 2026 in Artificial Intelligence
Strike 3 Holdings and Counterlife Media have sued Meta, accusing it of downloading thousands of adult films via BitTorrent to train AI models, which Meta denies. Meta's internal investigation has identified employees who appear to have interacted with the videos, though it says it has no evidence the films were torrented by them or anyone else.
Why it matters
The case puts a major tech company's data sourcing for AI training under legal scrutiny and tests how such claims are investigated and disclosed. The disputed download activity came largely from third-party IP addresses, which complicates attribution.
Meta's corporate IP addresses largely belong to an internet service it acquired years ago and phased out in 2019, and Onavo VPN was shut down in early 2019. Meta has named five employees in a document shared under an Attorney's Eyes Only designation, and says it has not yet identified the source of the torrenting activity.
The players
Meta
Meta is the technology company that owns Facebook, Instagram and WhatsApp and is the defendant in the lawsuit, which it denies.
Strike 3 Holdings
Strike 3 Holdings is an adult film producer that filed the lawsuit against Meta alongside Counterlife Media over alleged BitTorrent downloads.
Judge Eumi K. Lee
Judge Eumi K. Lee is a federal judge in the U.S. District Court handling the case, who denied Strike 3's request to relate the John Doe case without prejudice.
The details
Meta's internal search identified employees who appear to have interacted with the producers' videos, and it first shared five names in a July draft sent by mistake before clawing it back. Meta says the John Doe at issue had left the company before the alleged torrenting activity occurred.
Timeline
July 2026: Meta sent a draft with five employee names to Strike 3 by mistake.
October 15, 2026: A magistrate judge allowed disclosure of the John Doe's identity.
October 21, 2026: A joint status report was filed with large redactions.
October 28, 2026: Meta refiled with narrowed redactions.
The Tech Race
The lawsuit extends a wave of copyright litigation against AI developers over how training data is acquired from the internet. It highlights how the race to build larger AI models has pushed companies to scrutinize their own data-sourcing practices.
For users, the case raises questions about how the data behind AI tools is gathered and whether their own uploaded content could be used in similar ways. The outcome could influence how companies disclose data-sourcing practices in future products.
The takeaway
Companies building AI systems face growing legal scrutiny over where their training data comes from, and internal investigations may not settle questions about third-party activity. Readers should watch how courts handle disclosure and attribution in these disputes.
Further reading
For broader context on how AI companies source training material, see more on Artificial Intelligence.
Source note: This article includes information reported by Torrentfreak.
