Meta Pushes Back Against Porn Piracy Lawsuit, Calls AI-Training Claims “Nonsensical”

Meta is pushing back hard against a lawsuit accusing it of illegally downloading and distributing thousands of pornographic videos to train its artificial intelligence systems. In a motion filed Monday with the U.S. District Court for the Northern District of California, Meta asked the judge to dismiss the case entirely, calling the allegations “baseless” and “built on speculation.” The company insists there is no evidence whatsoever that its AI models were ever trained on copyrighted material.

“The Claims Are Built on Guesswork” According to Meta’s filing, the plaintiffs “desperately stitched together a story out of assumptions and coincidences,” yet failed to present any concrete proof. The company described the accusations as “unconvincing, inflated, and legally irrelevant.” The lawsuit was originally filed in July by Strike 3 Holdings, a Miami-based adult film distributor behind brands such as Vixen, Blacked, and Tushy.

Strike 3 alleged that Meta had used both corporate and “masked” IP addresses since 2018 to torrent nearly 2,400 adult films as part of a wider effort to train its multimodal AI systems. Meta has flatly denied the accusations.

“The Alleged Scale Makes No Sense” According to the court filing, only 157 videos were supposedly downloaded over seven years — an average of about 22 per year — across multiple IP addresses. Meta argues that this reflects disconnected activity by individuals for personal use, not a coordinated corporate operation. Meta’s attorney Angela L. Dunning further noted that Strike 3 wrongly attributed over 2,500 additional IP addresses to Meta, even though many of them don’t belong to the company — one, she said, is registered to a nonprofit organization in Hawaii.

AI Training and the Boundaries of Copyright Meta also stressed that it has no knowledge or evidence of any such downloads and gained no benefit from them. Monitoring every file moving through its vast global network, it said, would be “neither feasible nor legally required.” Experts say that while Meta’s defense might appear unconventional, it could be legally effective.

According to Dermot McGrath of venture capital firm Ryze Labs, admitting that any data had been used would open Meta up to questions about “fair use” and force it to reveal details of its internal training and auditing systems. “By denying that the data was ever used, Meta shields itself from deeper scrutiny,” McGrath told reporters. However, he warned that if courts accept this approach, it could create a dangerous precedent, allowing corporations to obscure how they train their AI models. “It could undermine copyright protection in AI training cases and make future oversight nearly impossible,” he added.

“Red Teams” and Legitimate Use of Explicit Data McGrath also acknowledged that processing explicit content isn’t automatically nefarious.

“Most major AI firms run red teams — specialized groups that test models with harmful or banned prompts to identify vulnerabilities,” he said.

“To create effective safety filters, you actually need to train the systems on examples of what you’re trying to block.” Still, Meta maintains that it had no knowledge of any pirated material and never used such data for AI development. The case now awaits a ruling from the federal court in California.

#meta , #AI , #technews , #ArtificialInteligence , #worldnews

Stay one step ahead – follow our profile and stay informed about everything important in the world of cryptocurrencies! Notice: ,The information and views presented in this article are intended solely for educational purposes and should not be taken as investment advice in any situation. The content of these pages should not be regarded as financial, investment, or any other form of advice. We caution that investing in cryptocurrencies can be risky and may lead to financial losses.“