The Tech Transparency Project, in findings shared with the Guardian, created four test accounts — two Apple, two Google — all indicating the user was 13. Researchers searched app stores using terms such as “AI video generator” and “adult AI” and said they were able to download 135 apps with features that could strip clothes from images or swap faces on to nude bodies. In their testing, the group said it used AI-generated women rather than photographs of real people.
The practical problem here is ugly and simple. These apps make it easier to turn ordinary photos into sexualized images without consent, and the Guardian linked the report to a Pennsylvania case involving Morgan Dorfman, who learned that a student at her daughter’s high school had used an app called Movely to create sexual videos using her daughter’s image.
Dorfman told the Guardian that the girls involved were “very scared and confused about what to do next”. She said she had known what a deepfake was in the abstract, but not that “it was possible to take any picture or any screenshot of anybody that you have and do this”.
What the report appears to show is that the weak point is not just the existence of these apps, but the way they are labeled and distributed. Apple and Google both prohibit this kind of app on their platforms, and both also use age ratings and parental controls as a gate. But the Tech Transparency Project said many of the apps it found were rated as suitable for minors, meaning they stayed available even when parental controls were switched on.
One face-swapping app, according to the report, was rated appropriate for ages four and up while being advertised with sexual imagery. Researchers also said a 13-year-old test account was able to buy a subscription to one app without any alert going to a linked parental account. Of the four accounts created for the test, two were placed under parental controls that allowed downloads only for apps marked suitable for 13-year-olds, with parental signoff. The researchers said neither Apple nor Google required or suggested that signoff for any of the nudification apps they accessed.
Google told the Guardian it regularly takes “proactive steps to detect and remove apps with harmful content” and said it had suspended violating apps and restricted related search terms after the findings were reported. Apple said it prohibits nudify apps on the App Store, removes developers that add those capabilities after review, and had taken action on the apps named in the report. It also said some apps cited were mainstream photo and video editing tools without nudification or deepfake capabilities.
That leaves the report’s narrowest finding looking like the most important one: parental controls did not reliably function as a barrier when app ratings and disclosures were wrong.
Alex de Valletta
Dua Mifsud
Sophia Borg