Deep Fakes of OnlyFans Stars Used in Scams

Elaina St. James was scrolling TikTok when she saw herself talking. Except she didn’t create that video. It was a deepfake video featuring an AI version of St. James that used a photo from her book launch photoshoot. The video was trying to drive viewers to follow their account instead of St. James’s real account. St. James, who has made more than $1.5 million on OnlyFans since 2021, had a hard time understanding what she was seeing. She knew she didn’t film this, but there she was. Then she noticed the eyebrows. They never moved, no matter what “her” mouth was doing, which is a dead giveaway for deepfakes. 

This is the new normal for creators in the adult content economy. Imposter accounts have existed as long as content platforms have, typically using stolen photos to create fake profiles. But the rise of AI tools has made it easier than ever for scammers to generate convincing video and voice from just a handful of source media. It’s happening at such speed and volume that the old-fashioned DMCA whack-a-mole response feels almost quaint. Jessieanna Campbell says fans regularly message her furious after sending $150 through CashApp to someone impersonating her on Snapchat, only to get blocked the second the payment clears. Jaime Smiles has scammers telling her fans that her daughter is in the hospital, fabricating a crisis specific enough that people believe it and send money. Shaghana Doyle has had fans who fell for impersonators show up at her actual home, convinced they have a real relationship with her after communicating with the impersonator for weeks.  

Creators are spending their own money to try to slow down the scams. Reba Rocket, co-CEO of Takedown Piracy, says her clients spend anywhere between $45 and $150 a month just ot keep pace with the content that used to require real Photoshop skill and now only takes one click to create dozens of variations. Her firm has removed nearly 400 million pieces of infringing content total. But the damage is more than just financial. Some creators’ images aren’t just being used for scams like these. They’re being grafted onto content they never filmed, watermarks and all, and hosted on servers in countries where takedown requests can be ignored. “As a sex worker, I know I’ve already signed up to sell videos and photos of myself in certain ways,” Doyle says. “But I didn’t sign up for this.”

That’s the point adult creators are reinforcing. Yes, they signed up to create adult content. But on their own terms. Campbell put it plainly: her job runs on consent, paperwork, boundary-setting, actual conversations about what gets filmed and shared. A deepfake skips every part of that process and keeps her face. “I didn’t consent for them to remove my bathing suit on a picture,” she says. “Yeah, I have nudity on the internet. I do porn. But this is my body.” Women make up 90 percent of victims of nonconsensual intimate imagery, and Doyle’s read is blunt: this isn’t just an attack on sex workers. It’s an attack on all women. Because the same tools are being used to take creators of totally PG content and use their images on adult content, too. 

There’s another layer to this problem beyond money. The fans showing up to try to meet Doyle show that the feeling of intimacy no longer requires you to be in the same room with someone – or even be a real person. She explained the real danger of a fan who’s built an entire imagined relationship with a fake version of her, meeting the real her, and reacting badly to the gap between the two. That’s what happens when a stranger’s sense of connection gets manufactured by a tool that doesn’t care whether the person on the other end consented to being loved by an audience they never met. 

For now, creators are doing triage with the tools available to them, like takedown services, watermarks, and disclaimers plastered across their bios warning fans that if it feels too easy, it’s not really them. Campbell’s outlook isn’t optimistic. What’s going on isn’t fair, but all they can do is fight and hope that detection keeps pace with generation.