WIRED found 400 apps and websites masking abusive deepfake features. They posed as pet portrait or language tuition tools. Hidden interfaces allowed explicit image and video generation. Templates included violent scenes like bedroom rape or knife threats. The network was linked to Chinese developers and KeLing Technology Limited. Apps were promoted via tens of thousands of Meta ads. Some ads contained AI-generated child sexual abuse material. The apps bypassed Apple and Google app store reviews. Developers changed content after approval to add explicit features. Meta banned KeLing Technology Limited from advertising in 2025. Apple removed 19 apps after WIRED’s inquiry. Google removed most apps before being contacted. Meta took action against 11,600 ad accounts and 4,700 business pages.
What changed
Apps appeared as benign AI tools like pet portrait generators or language tuition services.
What it unlocks
Users could access hidden interfaces to generate sexually abusive deepfake images and videos.
- 400 apps and websites
- 350,000+ downloads
- 100,000+ downloads for DiffAI
- 3,000+ associated ads
What you need to act on it
- clicking associated ads on Meta platforms
- server-side determination based on user data
Sources