Meta Denies Using Pornography To Train AI, Says Alleged Downloads Were ‘Personal Entertainment’ By Employees

Meta Faces Unusual Lawsuit Over Alleged Piracy Of Adult Films For AI Training

Meta is fighting a peculiar lawsuit that accuses the company of downloading thousands of adult films to train its artificial intelligence models — and in an ironic twist, it’s using the same legal defence once used by BitTorrent pirates.

Adult Studios Accuse Meta Of Massive Piracy

Adult film producers Strike 3 Holdings and Counterlife Media, known for brands such as Vixen, Tushy, Blacked, and Deeper, have sued Meta, claiming the company pirated at least 2,396 of their “award-winning, critically acclaimed adult motion pictures” since 2018.

They allege these videos were used to train Meta’s AI systems, including Meta Movie Gen and its large language model, LLaMA.

After Meta acknowledged in a separate case that it had obtained some material from unauthorised sources, Strike 3 said it combed through its own BitTorrent data and discovered IP addresses linked to Meta repeatedly downloading adult content.

The companies claim this activity was part of a broader effort to gather AI training material.

Meta Dismisses The Allegations As ‘Nonsensical’

Meta responded this week with a motion to dismiss the case, calling the allegations “nonsensical and unsupported.”

The company argued that the plaintiffs’ theory “relies on guesswork and innuendo,” adding that no facts show Meta ever trained an AI model using adult videos or images.

According to Meta, the timeline alone undermines the claims.

The alleged downloads began in 2018, years before Meta began its research into video-based AI models in 2022.

Meta stated in its filing,

“Plaintiffs go to great lengths to stitch this narrative together with guesswork and innuendo, but their claims are neither cogent nor supported by well-pleaded facts.”

Meta’s Pirate Defence Turns The Tables

In an unexpected twist, Meta has invoked the same legal reasoning that Strike 3 has long battled against in court.

The company argues that IP addresses alone cannot identify who downloaded the material — a position upheld in previous rulings by the Ninth Circuit Court of Appeals.

Meta labelled Strike 3 a “copyright troll,” referencing its reputation for filing thousands of lawsuits against alleged individual infringers.

It also questioned the plaintiffs’ logic, pointing out that if Meta had truly wanted to conceal piracy, it would not have used corporate IP addresses visible to the public.

Was It Really About AI Training Or Personal Entertainment?

Meta offered a more mundane explanation for the activity, suggesting that employees, contractors, or visitors at its facilities may have used the company’s network “for their own personal entertainment.”

The numbers appear to support this argument — roughly 22 downloads per year across numerous IP addresses — a figure Meta says is far too small to suggest a coordinated effort to amass data for AI training.

Meta noted, implying that the plaintiffs’ claims were implausible,

“Anyone familiar with AI development knows that training video models requires enormous datasets.”

No Grounds For Liability, Says Meta

Meta also rejected claims of contributory and vicarious copyright infringement, which would make the company responsible for others’ actions on its network.

For such claims to succeed, Meta would have to profit from the activity and have the ability to control it — conditions the company says do not apply.

Citing the Ninth Circuit’s Cobbler precedent, Meta argued that corporations are not required to monitor every download occurring on their networks.

The company said,

“Monitoring every file downloaded by any person using Meta’s global network would be an extraordinarily complex and invasive undertaking.”

Ongoing Legal Scrutiny Over AI Training Data

The case joins a growing list of copyright battles involving AI companies accused of using protected works as training data.

Meta itself has faced similar lawsuits in both the US and France from authors claiming their work was used without consent.

Strike 3 and Counterlife Media have two weeks to respond before the California federal court decides whether the lawsuit proceeds or ends here.

Regardless of the outcome, this case has quickly become one of the strangest intersections yet between adult entertainment, digital piracy, and artificial intelligence.