Apple and Google’s parental controls are failing to reliably prevent teenage users from accessing apps capable of producing nonconsensual sexualized imagery, according to a new investigation that raises questions about how app stores enforce age ratings and AI-content policies.
The Tech Transparency Project created test accounts representing 13-year-old users and was able to download 135 apps with features capable of generating nonconsensual nude imagery, The Guardian reported.
The findings come despite explicit policies from both Apple and Google restricting apps that create or facilitate nonconsensual sexual content.
Parental controls did not stop access
The investigation used four test accounts—two with Apple and two with Google—registered as belonging to teenagers.
Researchers found apps that could digitally remove clothing or place someone’s face onto sexualized bodies, even though parental-control settings were enabled. The Guardian reported that the researchers successfully downloaded 135 apps during the tests.
The issue is not theoretical for families.
The report highlighted Morgan Dorfman, a Pennsylvania mother whose daughter was among six girls allegedly targeted when another student used an AI app called Movely to create sexualized videos using their images.
That turns an app-store moderation failure into something much more consequential: software capable of generating abusive imagery can reach the same young people most likely to know one another personally through school and social networks.
Apple has tightened its rules
Apple updated its App Review Guidelines on October 16, explicitly stating that moderation requirements now cover nonconsensual intimate imagery including imagery generated using AI tools. Developers must proactively prevent and remove that material or risk violating Apple’s Developer Code of Conduct and being removed from the App Store.
Apple also clarified its prohibition on apps that hide functionality during review and later introduce capabilities such as nudification, a practice the company describes as bait-and-switch.
Those changes show that moderation increasingly has to evaluate what software can actually do after installation, not simply what developers claim during submission.
Google’s policies already prohibit the same behavior
Google Play’s AI-generated content rules list AI-generated non-consensual deepfake sexual material among examples of prohibited content.
Its broader inappropriate-content policy also explicitly prohibits apps that claim to undress people or see through clothing, including services labelled as entertainment or pranks.
The gap, therefore, is not simply missing policy.
It is enforcement.
App-store safety increasingly depends on capability testing
For product and trust-and-safety teams, the investigation exposes a difficult moderation problem.
Generative apps can present themselves as general-purpose image editors, video tools or face-swapping services while retaining capabilities that can be misused after installation. Age ratings and keyword filters alone may therefore fail to capture actual functionality.
That means platforms may need more aggressive capability testing, continuous post-approval monitoring and stronger connections between parental controls and rapidly changing generative features.
App stores have spent years positioning centralized review as an important layer of device security.
AI now makes that promise harder to keep.
The question is no longer only whether an app contains prohibited content when reviewers examine it. Platforms increasingly need to determine what the app is capable of generating once it reaches a user’s phone.
