American tech giant Meta has landed in an unusual controversy and bizarre allegation. A new lawsuit alleges that the company trained its AI models using thousands of illegally downloaded adult movies.
Meta’s response? In short, no, we didn’t… and if any adult videos were downloaded, it was probably someone watching them on office Wi-Fi for personal entertainment.
What’s the lawsuit about Adult film studio Strike 3 Holdings has accused Meta of downloading nearly 2,400 of its copyrighted videos through BitTorrent. It claims these videos were secretly used to train Meta’s AI models, including the video tool Movie Gen and the LLaMA language model. However, Meta argues the timeline itself doesn’t make sense. Strike 3 says downloads began in 2018, while Meta says it only started serious multimodal AI research in 2022, much later. Meta: “No facts to support this claim” In its court filing, Meta called the lawsuit “nonsensical and unsupported.” The company insisted there’s no proof linking these downloads to its AI training. Meta said: There are no facts to suggest that Meta has ever trained an AI model on adult images or video, much less intentionally so. Meta said the extremely low number of downloads, roughly 22 adult videos a year, doesn’t hint at a coordinated effort. “The far more plausible inference… is that disparate individuals downloaded adult videos for personal use,” Meta stated. The company also argued that tracking every single file downloaded on its massive global internal network is nearly impossible and would be “extraordinarily complex and invasive.” Meta fires back at Strike 3 Meta didn’t hold back in questioning Strike 3’s motives. It accused the studio of having a history of filing mass lawsuits to pressure settlements. Meta wrote: Strike 3’s narrative is stitched together with “guesswork and innuendo” and lacks “well-pleaded facts. The company says the lawsuit does not name who downloaded the videos, whether they worked at Meta, or if any of the content ever reached an AI training dataset. Meta’s stand on adult content in AI Meta has repeatedly said that it avoids adult material in its AI systems. A Meta spokesperson said: We don’t want this type of content, and we take deliberate steps to avoid training on this kind of material. Its policies also prohibit generating explicit content with its tools, which Meta says contradicts the idea that such material would be useful to train its models. Meta under scrutiny for AI content rules This lawsuit comes at a sensitive time for Meta. A recent Reuters report claimed that some Meta AI chatbots were allowed to: Following the backlash, Meta announced updates to its AI safety policies. What happens next Meanwhile, Meta recently won a major copyright case where a US court ruled that training LLaMA on books qualified as fair use. If the court sides with Meta again, this could become another precedent-setting moment in the ongoing “AI training vs copyright” debate, this time with a sensational twist involving adult content.
The post Meta denies training its AI models using pornographic material:Tech giant says staff may have downloaded videos for ‘personal use’ appeared first on Tri-Cities India.
