Hundreds of Mundane AI Apps Are Hiding Sexually Abusive Deepfake Features
wired ·
A WIRED investigation has uncovered a network of over 400 mobile applications and websites that disguise sexually abusive deepfake capabilities behind benign AI tools. These apps, often marketed as pet portrait generators or language tutors, secretly load explicit interfaces when users click on specific advertisements from Meta platforms. The operation appears to be run by Chinese developers linked to an entity named KeLing Technology Limited, who utilize deceptive 'bait-and-switch' tactics to bypass App Store reviews. Researchers found that the apps dynamically determine whether to show safe content or pornographic templates based on user data and ad source. Despite policies prohibiting such material, the network operated for approximately a year, accumulating hundreds of thousands of downloads before being partially removed by Apple and Google. Major tech companies like Meta have since taken action against thousands of associated ad accounts and domains following the exposure. Experts describe this as an organized commercial enterprise designed to evade enforcement through disposable assets and fake identities.
wired