WIRED Finds 400+ Deceptive AI Apps Hiding Sexual Deepfake Tools

A WIRED investigation found that more than 400 connected domains and apps that often present themselves as ordinary AI services can stealthily load interfaces for generating explicit AI sexual images and videos. Researchers discovered the network after investigating links in thousands of video ads with explicit deepfake content on Meta's Facebook and Instagram.

Deceptive apps and hidden behavior

One example, a pet portrait app called Viefur, appears to turn pet photos into anthropomorphic portraits. Researchers say apps in the network can make server-side determinations about whether to show benign content or load an explicit deepfake web interface, using information such as location, language and a unique identifier. Downloading an app after clicking an associated Meta ad appears to be one trigger. The apps include explicit AI scenes such as bedroom rape and threatening knife, according to WIRED.

Scale and platform response

Researchers found the apps had been published for around a year, at times dozens per week. Paul Bouchaud of AI Forensics found that around 30 nudify apps and websites were being registered each week at one point, with 27 live on Google Play in September. Those apps had been downloaded more than 350,000 times collectively, and many were rated suitable for all age groups. One app, DiffAI, had more than 100,000 downloads. Google removed DiffAI shortly before WIRED contacted the company and had removed most of the apps before being contacted. Apple removed 19 apps, including Viefur, after WIRED asked about the activity. Apple spokesperson Adam Dema said the apps complied when reviewed but developers later changed the website the app pulls content from. Meta said it banned KeLing Technology Limited from advertising in 2025 and later took action against thousands of ad accounts, pages and users linked to the domains.

Chinese origins and company links

WIRED reported that the apps and domains appear to be run by Chinese developers and a Hong Kong registered entity called KeLing Technology Limited. Maël Le Touz of Infoblox said the network uses exclusive Chinese hosting, code written in Chinese and comments in Chinese. The company's name was listed as contact details for the first website registered in the network, and an administration login page linked to the explicit deepfake interface mentions Keling. Email addresses linked to the entity, apps and related services did not respond to multiple requests for comment. After WIRED reached out, some public connections between KeLing and the deepfake tools appeared to be erased.

Sources: Wired