Technology news and discussion on TalkFuse.
Hundreds of Mundane AI Apps Are Hiding Sexually Abusive Deepfake Features
Researchers found a network of more than 400 apps and websites that present themselves as ordinary services, such as language tuition or AI fitness tools, but can secretly load tools for making sexually abusive deepfakes. The discovery matters because the apps were promoted through explicit adverts on Facebook and Instagram, while being distributed through Apple and Google app stores and allowing users to pay through their services.
The apps appear to choose what users see based on information such as location, language and whether they arrived through an advert; researchers say clicking a Meta advert may trigger the explicit version. The hidden interface offers dozens of sexual image and video templates, including scenes involving rape or threats with a knife, and lets users upload photos of people and buy “diamonds” to generate images. Researchers say the network has operated for about a year, sometimes releasing dozens of apps weekly, and appears linked to Chinese developers and Hong Kong company KeLing Technology Limited.
In short
- More than 400 apps and domains may conceal deepfake tools.
- The apps can switch between benign and explicit interfaces.
- Researchers link their promotion to explicit Meta adverts.
Explained simply
Deepfakes are images or videos altered or generated with artificial intelligence to make someone appear to do or say something they did not. When real people’s faces are placed into sexual material without their consent, this is often described as a sexually abusive deepfake.
The apps in this story are said to look like everyday tools, while showing different features to some users depending on details such as their location or how they found the app. Researchers have linked many of them to Chinese developers and a Hong Kong company called KeLing Technology Limited.
The case involves several parts of the online system: adverts on Facebook and Instagram, apps distributed through Apple and Google’s stores, and payments made through their services. It raises questions about how platforms identify and respond to abusive services promoted or sold through them.