Meta denies torrenting 2,400 porn movies for AI training – says they were for "personal use"

The company behind Vixen, Tushy, Blacked, and Deeper wants $359 million in damages

by · TechSpot

Serving tech enthusiasts for over 25 years.
TechSpot means tech analysis and advice you can trust.

Facepalm: After being accused of illegally torrenting pornography to train its AI systems, Meta is asking a court to throw out the case because the content was likely for "personal use." Meta also claims there is no evidence of it using adult images or video to train its AI systems.

We've seen several instances of AI companies being hit with lawsuits. Meta was served with another of these complaints in July, filed by Strike 3 Holdings and Counterlife Media, makers of popular adult entertainment brands such as Vixen, Tushy, Blacked, and Deeper.

According to TorrentFreak, the companies allege that Meta downloaded at least 2,369 of their movies since 2018 to train its AI.

The claim stems from a lawsuit filed against Meta by a group of authors in 2024. After that case revealed Meta's torrenting activity, Strike 3 Holdings checked its BitTorrent-tracking tools designed to detect copyright infringement of its videos. This revealed several corporate IP addresses identified as owned by Meta. It's further alleged that Meta concealed a "stealth network" of 2,500 "hidden IP addresses."

Why would Meta want to use almost 2,400 porn videos to train its AI? Strike 3 says it was to secretly train an unannounced adult version of its AI model powering Movie Gen. As such, the producer is demanding $359 million in damages.

Meta, of course, denies all the accusations and has asked for the suit to be dismissed. It noted that Strike has been labled by some as a copyright troll that files extortive lawsuits.

No porn here, honest

The social media giant has several arguments for its defense. It notes that 2018 – when Strike 3 says the infringements began – was years before Meta started even researching LLMs and generative video, which began in 2022.

// Related Stories

There's also the important fact that Meta's terms prohibit the generation of adult content by its AIs.

As for the IP addresses being used as evidence, Meta says that the small number of downloads – roughly 22 per year on average across dozens of Meta IP addresses – plainly indicates that the videos were torrented for personal use. In the authors' piracy case, for example, Meta admits to using the Book3 dataset, a 37GB compilation of 195,000 copyrighted books to train its LLMs.

With tens of thousands of employees and numerous contractors, visitors, and third parties accessing the internet at Meta every day, the company says it's impossible to know who downloaded the clips or if they were Meta employees.

There's also a claim that a Meta contractor was instructed to download adult content at his father's house. Meta argues there's no reason an automation engineer would be expected to source AI training material in his role, especially at his father's home, and that this was also obviously a personal use case.

"[T]hese claims fail not only for lack of supporting facts, but also because Plaintiffs' theory of liability makes no sense and cannot be reconciled with the facts they do plead. The entire complaint against Meta should be dismissed with prejudice," Meta concludes.

In August, Anthropic agreed to pay authors $1.5 billion over alleged use of pirated books for LLM training. Apple, meanwhile, was sued by authors in September over claims it used a known body of pirated books to train its OpenELM LLM.