Meta has pushed back against allegations that it illegally used pornographic material to train its AI models, claiming instead that the adult content may have been downloaded by a few of its employees for ‘personal use’.
In a legal filing, Meta asked a US district court to dismiss a copyright infringement lawsuit filed against the social media giant by Striker 3 Holdings, a company that says it makes high quality, ethical adult videos. The lawsuit accused Meta of using Striker 3’s copyrighted-protected porn videos to train an unannounced AI model that powers its video generator Movie Gen.
Denying these allegations, the Facebook parent said that there was no evidence that it directed employees to illegally torrent about 2,400 adult movies owned by Strike 3 for the purpose of AI training. There are “no facts to suggest that Meta has ever trained an AI model on adult images or video, much less intentionally so,” it argued.
The lawsuit states that Meta has been illegally downloading Striker 3’s movies since 2018, four years before Meta’s AI efforts “researching Multimodal Models and Generative Video” began. This makes it implausible that the downloads were intended for AI training, Meta argued.
The company further said that since activity on corporate IP addresses only showed about 22 downloads per year“the far more plausible inference to be drawn from such meager, uncoordinated activity is that disparate individuals downloaded adult videos for personal use,” Meta said.
Striker 3’s lawsuit against Meta has drawn attention in the context of the company’s internal rules that used to permit its AI chatbots to “engage a child in conversations that are romantic or sensual,” generate false medical information, and help users argue that Black people are “dumber than white people”, as per a Reuters report.
Meta has since announced changes to its policies. The company also reiterated in the legal filing that its terms prohibit users from generating adult content through its AI models. This is in contradiction to “the premise that such materials might even be useful for Meta’s AI training,” the company said.
Story continues below this ad
Strike 3 “does not identify any of the individuals who supposedly used these Meta IP addresses, allege that any were employed by Meta or had any role in AI training at Meta, or specify whether (and which) content allegedly downloaded was used to train any particular Meta model,” Meta wrote in the filing. “We don’t want this type of content, and we take deliberate steps to avoid training on this kind of material,” a company spokesperson was quoted as saying by ArsTechnica.
Additionally, the Strike 3 lawsuit alleges that Meta concealed using a “stealth network” of 2,500 “hidden IP addresses” to torrent the pornographic content. The lawsuit seeks damages above $350 million from Meta. Strike 3 reportedly has two weeks to respond to Meta’s motion, which argues that Strike 3 had failed to prove the company had anything to do with the illegal downloads.
Earlier this year, Meta scored a win in a major copyright case after a US district court sided with the company on the use of books to train its Llama models as being protected by the fair use doctrine in US copyright law.
