400 AI Apps Hide Abusive Deepfake Features

WIRED has uncovered roughly 400 apps and websites tied to Chinese-language developers that appear to be ordinary AI tools but contain hidden features for generating violent and sexually abusive deepfake images and video. The finding highlights a troubling pattern: abusive image-generation capabilities are being packaged inside mainstream-looking products, making them harder for users, app stores, and regulators to spot. On the surface, many of these services may look like familiar AI utilities, such as photo editors, avatar makers, or creative assistants. The hidden functions, according to the report, allow users to produce explicit and harmful synthetic media. That combination creates enforcement challenges because the abusive features may sit behind paywalls, invite codes, or other layers that are not obvious during a routine app review. App stores can remove products that violate policy, but the underlying websites and apps can reappear under new names or distribute outside official channels. Regulators face similar difficulties as tools move quickly across borders and platforms. The report also raises questions about how developers should be vetted and how quickly abusive capabilities can be detected before they reach a wide audience. For users, the findings underscore that a polished interface is not proof of safety. For platform operators, they suggest a need for deeper scrutiny of AI products, especially those that handle image and video generation. The case is part of a larger debate over how to curb deepfake abuse without crushing legitimate AI creativity. It also shows that detection and enforcement must keep pace with the ways harmful features are hidden inside seemingly normal software.