Meta claims that thousands of pirated adult videos it was accused of using for AI training may have been downloaded by 'disparate individuals' for 'personal use'
Usually, when adult videos are discovered on a device, the response is to claim some grander purpose for their existence. Well, Meta isn't one to bow to convention, apparently, as it has recently struck back at claims that adult videos were used to train its AI models, claiming that they may have been downloaded for 'personal use' instead.
In July of this year, a company called Strike 3 Holdings filed a suit against Meta, alleging that Meta had pirated "at least" 2,396 adult videos for use in its AI training. The original suit [PDF warning] filing claims that Meta, through BitTorrent, downloaded "award-winning, critically acclaimed adult motion pictures", referred to as Works, to "train its Meta Movie Gen, Large Language Model (“LLaMA”), as well as various other Meta AI Models that rely on video training content."
Meta filed a motion to dismiss (via Ars Technica) on October 27th, arguing these adult videos are "plainly indicative of private personal use, not a concerted effort to collect the massive datasets Plaintiffs allege are necessary for effective AI training."
This same motion argues that the downloads go back to 2018, averaging around 22 per year, and weren't substantiated by Strike 3 to be relevant to its data set. It says the "Plaintiffs’ supposition that Meta must have instigated these downloads for AI training is implausible on its face. All [the] Plaintiffs have are IP addresses, which is insufficient to state a claim."
Meta also claims that "tens of thousands of employees and innumerable contractors, visitors, and third parties access the internet at Meta every day. While it is possible one or more Meta employees downloaded Plaintiffs’ videos, it is just as possible…that a 'guest, or freeloader,' or contractor, or vendor, or repair person—or any combination of such persons—was responsible for that activity."
Effectively, while Strike 3 may have Meta IP addresses logged, Meta believes it can't claim any more than that. "Plaintiffs simply have no idea who downloaded these Works, and do not plead facts plausibly showing that it was Meta." The document also argues that, despite seemingly having the ability to find the owner of an IP address, Strike 3 fails to do so, "or, if they did, chose not to share the results of their findings in the Complaint."
Finishing off the motion to dismiss, Meta's lawyer says, "these claims fail not only for lack of supporting facts, but also because Plaintiffs’ theory of liability makes no sense and cannot be reconciled with the facts they do plead. The entire complaint against Meta should be dismissed with prejudice." Finally, Meta argues that there's no evidence that Meta encouraged its employees or contractors to download any adult content for any reason.
This case is complicated somewhat by the reputation around Strike 3, the plaintiff. Just a few months ago, copyright lawyer and YouTuber Leonard French called Strike 3 "the most litigious copyright troll" in regard to its lawsuit against Meta. According to Torrentfreak, Strike 3 has two weeks to respond to the motion, in which Meta may be allowed a follow-up response. Then, the motion will go forward to deliberation in the California federal court.