Strike 3 Holdings just dropped a bombshell lawsuit against Meta, alleging the tech giant has been torrenting and seeding thousands of copyrighted adult videos since 2018 to train its AI models. The $350 million federal case, with newly unsealed details, could set a massive precedent for how AI companies source training data - and it's already exposing the murky world of corporate data scraping behind the AI boom.
The adult entertainment industry just fired its biggest shot yet in the AI copyright wars. Strike 3 Holdings, which produces what it calls 'feminist' and 'ethical' adult content, is demanding $350 million from Meta in a federal lawsuit that paints a damning picture of how the social media giant allegedly built its AI empire.
According to court documents unsealed last week, Meta didn't just passively scrape Strike 3's content - it actively torrented and seeded 2,396 copyrighted videos since 2018, using BitTorrent protocols typically associated with piracy. The company's infringement detection systems tracked the alleged violations across 47 distinct Meta-affiliated IP addresses.
'They have an interest in getting our content because it can give them a competitive advantage for the quality, fluidity, and humanity of the AI,' Strike 3 attorney Christian Waugh told reporters. The lawsuit argues Meta specifically targeted adult content because it contains visual angles, body positioning, and extended uninterrupted scenes that mainstream movies and TV rarely provide - exactly what's needed to train sophisticated video AI models.
But the adult content was just the tip of the iceberg. Exhibits in the case reveal Meta allegedly torrented everything from Yellowstone and Modern Family episodes to weapons training videos like '3D Gun Print' and 'Gun Digest Shooter's Guide to the AR-15.' The list even includes content with potentially underage performers and, ironically, a video called 'Intellectual Property Rights in Cyberspace.'
The timing couldn't be worse for Meta. Just this week at Meta Connect, Mark Zuckerberg doubled down on the company's 'superintelligence' vision, promising to put AI power directly into users' hands through products like the company's smart glasses. Meta's V-JEPA 2 'world model,' released in June, was trained on one million hours of what researchers vaguely called 'internet video' - a term the lawsuit argues was deliberately unspecific.
'Using pirated material as training data was reportedly an active choice made by Meta executives and approved by Zuckerberg directly,' according to previous reporting. This aligns with a pattern across the industry - nearly every major AI company now faces similar copyright infringement allegations.
For Meta, the Stakes couldn't be higher. Matthew Sag, an AI law professor at Emory University, warns this is 'a public relations disaster waiting to happen.' He painted a nightmare scenario: 'Imagine a middle school student asks a Meta AI model for a video about pizza delivery, and before you know it, it's porn.'
The case could also become a legal watershed moment. In June, U.S. District Judge Vince Chhabria ruled in favor of Meta in a separate copyright case involving authors' books, but explicitly stated his ruling didn't mean Meta's AI training practices were lawful - just that those particular plaintiffs 'made the wrong arguments.'
That leaves room for Strike 3 to build a stronger case. 'The best version of their argument is: this is a fundamental problem because by going to these pirate websites, you are undermining the market for access,' Sag explained. The company claims to have detected a 'wide selection' of additional copyrighted content that Meta allegedly torrented, with the current exhibits representing only 'a thin slice of the pie.'
Meta spokesperson Christopher Sgro told reporters the company is 'reviewing the complaint, but we don't believe Strike's claims are accurate.' But the broader industry is watching nervously. Even President Trump has weighed in, arguing in July that 'You can't be expected to have a successful AI program when every single article, book, or anything else that you've read or studied, you're supposed to pay for.'
Waugh sees this as a defining moment for digital rights. 'It doesn't matter if it's a four-sentence poem or adult entertainment,' he said. 'There is no appetite in this country for what AI companies appear to be doing, which is making money off the backs of rights holders who never gave permission for it.'
This lawsuit represents more than just another copyright battle - it's a stress test for the entire AI industry's data sourcing practices. With Strike 3 claiming to have evidence of systematic piracy and nearly every major AI company facing similar allegations, the outcome could reshape how artificial intelligence gets trained. For Meta, which just promised to put 'superintelligence' in everyone's hands, the question isn't just legal liability - it's whether the company can deliver on those grand promises while the foundation of its training data faces unprecedented scrutiny.
