Meta finds itself in an unusual legal predicament, facing accusations from adult film producer Strike 3 Holdings that the tech giant pirated thousands of adult movies to train its AI systems. The company’s response? A spirited defense that borrows tactics from the very BitTorrent pirates Strike 3 has spent years pursuing.
Strike 3 Holdings, along with Counterlife Media, the companies behind popular adult entertainment brands like Vixen, Tushy, Blacked, and Deeper, filed a lawsuit claiming that Meta downloaded at least 2,396 of their films since 2018.
The damages they’re seeking are staggering: $359 million. Their case rests on evidence that forty-seven IP addresses owned by Facebook allegedly shared their copyrighted content through BitTorrent networks.
The adult film producers discovered these alleged infringements after Meta’s BitTorrent activity came to light in a separate lawsuit filed by book authors.
Meta Uses Pirate Defense Against Strike 3, Claims ‘Personal Entertainment’ Downloads
When Meta admitted in that case to obtaining content from pirate sources, Strike 3 and Counterlife Media decided to dig through their own archives of BitTorrent data. That’s when they found IP addresses linked to Meta repeatedly downloading their content.
This week, Meta fired back with a motion to dismiss, and the irony is hard to miss. The tech company is using the exact same defense that countless BitTorrent pirates have employed when facing Strike 3 in court.
Meta argues that IP addresses alone don’t prove who actually committed the infringement, a position the Ninth Circuit Court of Appeals has previously supported.
Meta labels Strike 3 as a “copyright troll” and points out that rightsholders need “something more” than just IP address evidence to prove their case. The company insists there’s no proof that these downloads were part of any centrally orchestrated effort to gather training data for AI models.
The timeline is particularly problematic for the plaintiffs’ theory, according to Meta. The alleged infringing activity started in 2018, but Meta didn’t begin researching video models until 2022. How could the company have been downloading content for AI training purposes years before those AI projects even existed?
Rather than just denying the allegations, Meta offers an alternative explanation that’s both mundane and plausible: employees or visitors to Meta’s facilities probably downloaded the videos for their own personal entertainment.
The numbers support this theory, Meta argues. With roughly 22 downloads per year spread across dozens of IP addresses, the scale seems more consistent with individual users satisfying their personal interests rather than a corporation gathering massive datasets for AI training.
Anyone familiar with AI development knows that training video models requires enormous amounts of data, far more than what Strike 3 claims to have found.
Meta Seeks Dismissal of “Nonsensical” AI Training Copyright Lawsuit by Adult Producers
Meta also pokes holes in another part of the plaintiffs’ argument. Strike 3 suggested that thousands of IP addresses outside Meta’s network were used to conceal the company’s BitTorrent activities.
Meta counters with a logical question: Why would the company try to hide some downloads while using easily traceable corporate IP addresses for hundreds of others? The inconsistency, Meta argues, proves there was no coordinated scheme.
Meta doesn’t completely rule out that someone on its network downloaded these videos. However, the company argues it can’t be held liable for contributory or vicarious copyright infringement, legal concepts that would make Meta responsible for the actions of others using its network.
For vicarious infringement to stick, Meta would need to have a financial interest in the downloads and the ability to supervise or stop them. Meta argues it has neither.
The company cites the Ninth Circuit’s “Cobbler” case as precedent, which established that companies aren’t required to police every possible infringement on their networks.
Similarly, the contributory infringement claim fails because Meta says it had no knowledge of the pirating activity and didn’t materially contribute to it.
Meta is asking the court to dismiss the entire complaint with prejudice, meaning Strike 3 couldn’t refile the case. The company describes the adult producers’ AI training theory as “nonsensical” and argues that the claims lack supporting facts.
Strike 3 Holdings and Counterlife Media have two weeks to respond to Meta’s motion. After Meta files its follow-up response, a California federal court will decide whether this unusual case moves forward or ends here. Whatever the outcome, it’s certainly one of the stranger chapters in the ongoing legal battles over AI training data.
