An investigation found a network of more than 400 connected domains and apps linked to Chinese-language developers that offered tools to create sexually explicit AI deepfake images and videos. The apps appeared in the Apple App Store and Google Play Store while presenting themselves as harmless services such as pet portrait generators, language tutors, and AI workout coaches.
Behind that front, users could access a hidden web interface to upload photos and generate explicit deepfakes. Some templates were labeled “bedroom rape,” and one depicted a man holding a knife to a woman’s neck. Many users reached the apps through explicit deepfake ads on Meta’s Facebook and Instagram platforms, where they could then pay through Apple Pay or Google Pay.
Researchers said the network had operated for about a year and at times published dozens of new apps per week. Paul Bouchaud, lead researcher at AI Forensics, called it an “organized, commercially run operation.” Maël Le Touz, a staff threat researcher at Infoblox, identified almost 500 linked domains. As of September, 27 apps were still live on Google Play, with more than 350,000 combined downloads. One app, DiffAI, had more than 100,000 downloads before it was removed.
Apple removed 19 apps after receiving a list of domains and said it has suspended tens of thousands of apps over the past year for “bait-and-switch” tactics while updating App Store rules to restrict nudify tools. Google had removed most of the apps before it was contacted and did not respond to a request for comment.
The investigation grew out of a broader review of thousands of explicit deepfake ads on Meta’s platforms. Some ads included AI-generated child sexual abuse material and images of real children.
