Arizona women sue AI ModelForge over porn avatars generated from photos of real women

Three women in Arizona have filed a lawsuit against a group of men who, according to the plaintiffs, turned social media photos into AI porn avatars and sold not only the output but also instructions for producing it. The case matters not because of the size of the accounts involved, but because the alleged victims were not celebrities. According to the lawsuit, they were ordinary users with public profiles.

How the scheme worked

One of the plaintiffs, identified in court filings as MG, said she ran an ordinary Instagram account where she posted stories, photos with friends, and everyday moments. Over the summer, someone sent her a link to videos of a woman who looked very much like her: the same face, a similar body, and tattoos in the same places. She said the material looked convincing enough that strangers could mistake it for real.

The lawsuit alleges that three men from Phoenix, along with other unidentified participants, searched the internet for photos of young women and then used them to create fictional AI models. According to the plaintiffs, this content was posted on Instagram and TikTok and monetized on Fanvue. At the same time, a $24.95-a-month subscription sold through the Whop platform gave access to courses that taught other users how to build the same kinds of accounts.

The lawsuit describes what amounts to an assembly line:

• finding women with relatively small social media accounts
• collecting their photos and uploading them to CreatorCore to train a model
• using a separate app to "undress" the images and generate explicit content
• posting AI videos and photos on social media to build reach
• selling instructions and templates to new members through a paid community

"It was disgusting on every level," one of the plaintiffs said of what she saw.

Why the lawsuit stands out

This case differs from many deepfake porn stories because it involves not only publishing images without consent, but also selling a ready-made playbook. The plaintiffs' attorney claims subscribers were allegedly taught how to choose women who would not be able to defend themselves quickly, and where to get their photos. The lawsuit specifically notes a preference for accounts with fewer than 50 thousand followers, to reduce the risk of legal claims.

According to figures cited in the lawsuit, the content built around this scheme racked up millions of views and brought in more than $50 thousand in revenue in a single month. The filing also states that in 2025 CreatorCore had more than 8 thousand subscribers, who generated over 500 thousand images and videos. Even if some of these numbers have yet to be verified in court, they show this is not a fringe story but a full-fledged economy built around AI influencers.

Another detail is the project's apparent rebranding. At the time of publication, AI ModelForge's Linktree pointed to TaviraLabs, a Telegram community with more than 18 thousand members that billed itself as the largest community for training AI-influencer models. This is typical for such schemes: changing the name, the platform, and the packaging lets operators quickly steer their audience away from negative attention without breaking the sales funnel or abandoning the old business model. It also makes it harder for both victims and platforms to connect new accounts to old ones.

Where protection breaks down

On paper, victims already have legal tools, but in practice those tools work slowly. The federal Take It Down Act was signed in May 2025 but takes effect only in May 2026. It is supposed to make publishing sexualized AI content without consent illegal and require platforms to remove such material within 48 hours of a complaint. Several states, including Arizona, have already passed similar bans, but local politicians themselves admit these measures more often respond to harm after the fact than prevent it.

Platform moderation is another problem. According to the plaintiffs, many of the images stayed up on Instagram because they did not technically fall under impersonation rules: they were not exact copies of their photos, but new AI-generated images with the same face and recognizable physical details. TikTok reportedly removed some of the accounts after journalists reached out, and Instagram sent the related profiles for review. But the gap between "not her photo" and "obviously made from her" remains a huge gray area.

What it means

The Arizona case shows that the AI influencer market has long since moved beyond harmless virtual models. If the allegations are proven, it will be an example of how generative tools, paid communities, and weak moderation combine into a scalable scheme for exploiting other people's likenesses without consent — and of how vulnerable almost anyone with a public social media profile can be. For platforms and lawmakers, it is a signal that the old model of complaints and manual takedowns can no longer keep up.

*Meta is designated an extremist organization and banned in the Russian Federation.