Meta has responded to a lawsuit claiming it illegally torrented pornography to train its AI model, saying the downloads were for "personal use" and not for AI training.
The tech giant filed a motion to dismiss the lawsuit on Monday, accusing the plaintiff, Strike 3 Holdings, of relying on "guesswork and innuendo." Meta claimed that there was no evidence that it directed any of the downloads or was even aware of the illegal activity.
According to Meta, the flagged downloads, which spanned seven years, started in 2018 - four years before its AI efforts began. This made it implausible that the downloads were intended for AI training, Meta argued. Furthermore, the company's terms prohibit generating adult content, contradicting the premise that such materials might even be useful for its AI training.
Meta's spokesperson described the claims as "bogus" and suggested that the alleged activity could have been conducted by a guest or contractor using one of Meta's global network IP addresses without any connection to the company. The spokesperson also pointed out that tens of thousands of employees, innumerable contractors, visitors, and third parties access the Internet at Meta every day, making it possible that a "guest" was responsible for the activity.
The lawsuit had claimed that Meta pirated adult films to secretly train an unannounced adult version of its AI model powering Movie Gen. Strike 3 had sought damages exceeding $350 million.
Meta's argument is that there was no evidence to support the claims, and the company has taken steps to prevent such content from being used in its AI models.
The tech giant filed a motion to dismiss the lawsuit on Monday, accusing the plaintiff, Strike 3 Holdings, of relying on "guesswork and innuendo." Meta claimed that there was no evidence that it directed any of the downloads or was even aware of the illegal activity.
According to Meta, the flagged downloads, which spanned seven years, started in 2018 - four years before its AI efforts began. This made it implausible that the downloads were intended for AI training, Meta argued. Furthermore, the company's terms prohibit generating adult content, contradicting the premise that such materials might even be useful for its AI training.
Meta's spokesperson described the claims as "bogus" and suggested that the alleged activity could have been conducted by a guest or contractor using one of Meta's global network IP addresses without any connection to the company. The spokesperson also pointed out that tens of thousands of employees, innumerable contractors, visitors, and third parties access the Internet at Meta every day, making it possible that a "guest" was responsible for the activity.
The lawsuit had claimed that Meta pirated adult films to secretly train an unannounced adult version of its AI model powering Movie Gen. Strike 3 had sought damages exceeding $350 million.
Meta's argument is that there was no evidence to support the claims, and the company has taken steps to prevent such content from being used in its AI models.
. Meta saying they didn't know about it and that it might've been done by some random contractor is pretty convenient, you know? And $350 million is a crazy amount of money 
. I mean, I'm all for protecting AI models from adult content, but the way Meta is handling this lawsuit seems kinda shady 
. They're basically saying "we had nothing to do with it" without even having to dig into the matter that much... it just feels like they're trying to sweep this under the rug 
. And what's up with Strike 3 Holdings' claims being based on "guesswork and innuendo"? That doesn't sound very solid 
I mean, I get it, who doesn't want a sweet AI that can generate adult content 
, but come on Meta! 
You gotta prioritize those ethics and safety protocols. 
, they're really owning this and saying "hey, we had no idea" 
which is kinda funny considering all the IP addresses and employees involved 
.
And what's with the $350 million damage claim? 
seems like a lot to me. Anyways, this is def a wild story and I'll be keeping an eye on it 
like who does that kinda thing anyway? and the lawsuit is asking for over $350 million, thats crazy! its clear that whoever filed that claim has some serious evidence but meta's trying to dismiss it out of hand. doesnt seem right though 
? Like, okay sure Meta but how come the downloads started happening four years before your AI thing even existed? It's a bit suspicious if you ask me 
. I'm not saying they definitely did it but come on Meta let's be real about this 
.
what's worse, though, is that some of those adult films might've been used in their AI model even tho meta claims to have taken steps to prevent it 