WIRED: Over 400 AI Apps and Websites Accused of Secretly Embedding Pornographic Deepfake Features

A WIRED investigation on November 2 uncovered a network of over 400 affiliated apps and websites that, under certain conditions, loaded pornographic deepfake interfaces disguised as ordinary AI tools for tasks like pet avatars and language learning, allowing users to upload photos of others to generate explicit images or videos. Researchers identified the connections through code, domain names, and advertising tracking. As of September, 27 of these apps were still available on Google Play, with a combined total of over 350,000 downloads, many of which were labeled as suitable for all ages.

After receiving tips from WIRED, Apple removed 19 of the apps, stating that developers had replaced the content with remote webpages after passing the review process. Google had already removed most of them before being contacted. Meta reported that it had blocked ads for 209 related apps. Based on clues such as Chinese-language code, hosting information, and domain names, the report linked the network to Chinese developers and a company registered in Hong Kong. The operators involved did not respond to requests for comment; this connection represents the investigative judgment of the media and researchers.

Source: WIRED