WIRED traces 400+ AI apps hiding abusive deepfake tools

A WIRED investigation traced a network of more than 400 connected apps and web domains that presented as pet-photo generators, language classes, or AI workout coaches while hiding server-triggered tools for making explicit deepfakes. Scenes available on the hidden interfaces included 'bedroom rape' and 'threatening knife,' where a man holds a knife to a woman's neck.

The apps passed review by behaving normally during submission. A server check could then activate the hidden features, and downloading the app after tapping an ad for the deepfake functionality on Meta platforms was one of the triggers. In September, 27 identified apps were live on Google Play with more than 350,000 downloads combined, many carrying all-ages ratings. The apps began appearing in late 2025, and their numbers rose sharply in spring 2026, with around 30 websites and apps for creating nude images registered each week at one point.

Apple removed 19 apps, including the pet-photo app Viefur, after reporters shared details. On October 16, Apple updated its guidelines to tighten restrictions on apps that conceal functionality during review, with a new Section 1.2 covering nonconsensual intimate imagery and guideline 2.5.19 covering hidden, dormant, or undocumented features. Meta said it acted against 11,600 ad accounts and 15,000 pages linked to the network.

Researchers tied the infrastructure to Chinese developers and KeLing Technology Limited, a company registered in Hong Kong, pointing to Chinese hosting, comments in Chinese in the code, and watermarks from Chinese video-editing software on promotional materials.

Two researchers we track posted the piece the day it ran.