Meta is facing a new lawsuit from Strike 3 Holdings, an adult film producer that says the company pirated more than 2,300 of its movies through BitTorrent to train AI models. The complaint alleges the downloads happened over several years, from 2018 through 2025, using IP addresses tied to Meta offices. The plaintiff says the pattern suggests the material was gathered to help develop advanced video capabilities.
The case builds on internal Meta documents revealed in February 2024, which indicated employees had used torrents to download large amounts of content for model training, including books. Strike 3 argues the same method was used here for adult content as well.
When Meta tried to have the case dismissed, it told the court the claims were baseless and said that if the downloads occurred, they were for “personal use” rather than for any official company activity or AI project. U.S. District Judge Aomi K. Lee was not persuaded. In a decision issued on June 11, she rejected Meta’s request to throw out the lawsuit and said the plaintiffs had provided enough evidence to move forward.
Lee pointed to the downloading patterns described in the filings, saying IP addresses linked to Meta had downloaded many similarly characterized files in large volumes on the same days. She said it was difficult to view that as coincidence or random activity by a few users. The lawsuit will now proceed, and both sides will have to present more evidence about where the downloads came from and whether the content was used to train AI systems. Strike 3 is seeking damages that could reach $359 million.
