How A.I. Is Making Abuse Easier

From School Hallways to City Streets, A.I. Tools Are Redrawing the Map of Abuse

A growing body of evidence is turning a once-theoretical debate over artificial intelligence into something far more immediate and intimate: a question of how easily new consumer technologies can be used to humiliate, expose, identify and pursue people without their consent.

New reporting on the spread of A.I.-generated nude images in schools, alongside a broad civil-liberties warning about the prospect of facial-recognition smart glasses, has sharpened concerns among educators, privacy advocates and lawmakers that the next phase of A.I. harm is not simply misinformation or academic cheating. It is exploitation.

In one arena, students — often girls — are discovering that ordinary photos can be transformed into fake nude images and circulated among peers in a matter of minutes. In another, advocates fear that smart glasses equipped with facial recognition could allow wearers to identify strangers silently in public, creating new tools for stalking, harassment and coercion.

Taken together, the developments suggest that A.I.’s most socially destabilizing effects may be arriving not through spectacular technological breakthroughs, but through the cheapening of abuse: making it faster, easier and harder to detect.

A School Crisis Larger Than Many Realized

The scale of the deepfake-nudes problem in schools appears to be broader than many parents and administrators had understood. Recent reporting found nearly 90 schools and about 600 students around the world had been affected by A.I.-generated fake nude images, a figure that almost certainly understates the problem because many cases go unreported.

Those numbers add to a stream of criminal cases that have made clear how quickly such abuse can spread among minors. In Pennsylvania, two teenagers were sentenced on March 25 after admitting to creating about 350 fake nude images of at least 59 girls, according to recent accounts of the case.

Child-safety researchers say the pattern is no longer anecdotal. Thorn, a nonprofit focused on online child sexual exploitation, found in a 2025 survey of 1,200 young people that 10 percent of teens personally knew someone who had been targeted by deepfake nudes, and 6 percent said they had been victimized themselves.

At the same time, reports involving A.I.-generated child sexual-abuse imagery have surged. The National Center for Missing and Exploited Children received 4,700 such reports in 2023; by just the first half of 2025, that number had ballooned to 440,000.

The rise has exposed a grim reality for schools: many were built to handle bullying, sexting and nonconsensual photo sharing, but not the speed and plausibility of synthetic sexual imagery made from innocent pictures. A student no longer needs a real explicit photograph to inflict lasting damage. A yearbook picture, a social media post or a snapshot from a sports event may be enough.

The New Logic of Harassment

What makes these tools especially alarming to advocates is not merely their sophistication, but their accessibility. “Nudify” apps and image-generation tools have lowered the technical barrier for abuse so dramatically that teenagers with little expertise can produce sexually explicit fakes of classmates. The resulting images can then circulate in the same social ecosystems — group chats, school networks, ephemeral messaging apps — where reputational harm spreads fastest.

For victims, the consequences are often immediate and profound: shame, social isolation, anxiety and the burden of proving that an image is fake after others have already seen it. Educators and prosecutors have struggled to catch up, in part because the conduct often falls between older categories of discipline and criminal law.

Congress moved in 2025 to address at least part of the problem. The TAKE IT DOWN Act, signed into law on May 19 of that year, created federal penalties and takedown obligations for nonconsensual intimate imagery, including A.I.-generated deepfakes. But the existence of a law does not guarantee swift enforcement, nor does it answer practical questions facing schools and families: whether platforms will remove such images quickly, whether app stores will curb tools designed for abuse and whether local authorities will treat peer-to-peer synthetic sexual exploitation with the urgency it demands.

A Warning About Smart Glasses

At the same time, another fight is unfolding over the next possible frontier for consumer A.I.: wearable facial recognition.

On April 13, a coalition of 75 civil-liberties and advocacy organizations, including the American Civil Liberties Union, the Electronic Privacy Information Center and Fight for the Future, urged Meta not to add facial recognition to its Ray-Ban and Oakley smart glasses. The groups warned that a reported feature known as “Name Tag” could allow users to identify strangers in real time, without their knowledge, by pairing camera-equipped glasses with biometric lookup tools.

The organizations argued that such a capability could be especially dangerous for abuse survivors, women, immigrants, political demonstrators and LGBTQ+ people — groups for whom anonymity in public can be a form of safety. In their view, facial recognition embedded in mainstream eyewear would not simply be another convenience feature. It would be a portable identification system that could be used in clinics, workplaces, protests, bars, sidewalks or school zones.

Meta said on April 13 that it does not currently offer facial recognition in those glasses and noted that competitors sell products with similar capabilities. But the company’s response did not quiet concerns, in part because lawmakers and advocates have been pressing the company for months about the possibility.

On March 17, Senators Ed Markey, Ron Wyden and Jeff Merkley demanded answers from Meta about whether biometric data would be deleted, whether it could be shared with law enforcement and what safeguards would prevent stalking or harassment.

The uncertainty is itself part of the debate. It remains unclear whether Meta will ship such a feature, how it would work or what restrictions might govern it. But privacy advocates say that waiting until mass adoption arrives would repeat an old mistake: allowing a new surveillance tool to become normalized before the public fully understands its risks.

When Manipulation Meets Identification

For years, discussion of A.I. harms often centered on broad concepts like misinformation, bias and trust. Those concerns remain. But the current wave of anxiety reflects something more concrete: the convergence of tools that can falsify a person’s body and tools that can reveal a person’s identity.

One technology can create intimate images that never existed. Another could potentially identify a stranger in seconds. Combined with ubiquitous cameras, social media archives and location data, those systems could deepen existing patterns of gender-based violence, stalking and social control.

Advocates say the danger is not hypothetical. A predator who can identify someone discreetly in public, search for their online presence and generate sexualized fake imagery or threats using readily available software would face far fewer barriers than before. The cost of intimidation drops; the speed increases.

That prospect helps explain why the debate has intensified now. The underlying technologies are no longer confined to laboratories or niche surveillance firms. They are appearing in school incidents, consumer apps and wearable devices sold as lifestyle products.

The Law Is Catching Up Slowly

The United States and other governments have begun to respond, but unevenly. New laws targeting nonconsensual intimate imagery have offered victims more avenues for removal and prosecution, and some school districts have updated codes of conduct to address A.I.-generated sexual abuse. Yet enforcement remains patchy, and many institutions still lack clear protocols for supporting victims, preserving evidence and punishing perpetrators.

The same lag is apparent in biometric privacy. Existing privacy, child-safety and civil-rights rules were not written for an era in which a pair of fashionable glasses might one day identify passers-by, nor for one in which a teenager can use a phone app to fabricate convincing explicit images of classmates.

The result is a widening gap between what consumer technologies make possible and what social institutions are prepared to stop.

For families, schools and policymakers, the stakes are becoming harder to dismiss. The question is no longer whether A.I. can be misused. It is whether the rules, products and platforms now taking shape will make exploitation rarer — or routine.

Sources

Further reading and reporting used to add context:

AI News

How A.I. Is Making Abuse Easier

From School Hallways to City Streets, A.I. Tools Are Redrawing the Map of Abuse