Nearly 90 schools. Around 600 students. Those are the documented cases from a joint investigation by WIRED and Indicator - and experts say they represent a fraction of the actual problem.
The investigation published by WIRED tracked confirmed incidents where AI-generated nude images of students were created and distributed within school communities around the world. The images aren't produced by sophisticated hackers. They're generated by consumer apps - often called "nudify" tools - that use AI image models to strip clothing from photos of real people. A student with a smartphone and a social media photo can produce one in minutes.
Why Schools Are Struggling to Respond
The legal and administrative framework in most countries wasn't built for this. In many jurisdictions, creating AI-generated nude imagery of a real person sits in a gray area, particularly when both the creator and subject are minors. Schools often find themselves first-responders to a crisis with no clear protocol: is it a disciplinary matter, a criminal one, or both?
Victims face compounded harm. The images are fake, but the humiliation is real, and they can spread across group chats and social platforms before anyone in authority knows they exist. Platforms hosting the content may take days to respond to takedown requests, if they respond at all.
The Apps Themselves Are the Infrastructure
Most nudify apps operate openly, marketing themselves as entertainment tools or AI art generators. Some run through browser-based interfaces requiring no download whatsoever. Despite growing legislative pressure, the market keeps expanding - new apps appear as older ones get taken down.
This is where the broader AI tooling industry has a direct stake. The same image models that power legitimate creative tools used by designers and marketers daily can be adapted (fine-tuned, meaning retrained from a general model to a specific task) for harmful applications. The distinction between a legitimate image platform and a nudify app often comes down to what guardrails the developer chooses to build.
Several AI image tools have introduced content filters under public pressure. DALL-E 3 includes filters designed to block the generation of nude images of identifiable real people. But enforcement is imperfect, and specialized apps exist specifically to work around these protections.
Who's Supposed to Fix This
The investigation surfaces a policy gap that legislators are slowly starting to address. In the US, several states have passed laws criminalizing non-consensual intimate deepfakes, and federal legislation has been proposed. The UK added similar provisions to its Online Safety Act. But enforcement is inconsistent, and cross-border cases - where the image creator and victim are in different countries - fall into jurisdictional gaps that prosecutors rarely have resources to pursue.
For schools specifically, the more immediate need is clearer guidance on how to respond when an incident happens: how to document it, who to notify, how to support victims, and how to handle perpetrators who are often also minors.
The 600 students in this investigation are confirmed victims. The actual number affected globally is almost certainly higher - most incidents go unreported because victims fear additional exposure or don't believe anyone can help.