There is an old spiritual instinct in nearly every tradition that warns against creating a false image of a person: a golem, a graven idol, a simulacrum that steals the likeness of a soul without its consent. Artificial intelligence has turned that ancient warning into a Tuesday-afternoon crisis in American middle and high schools. So-called “nudify” apps, which use generative AI to strip clothing from ordinary photographs, have moved from dark corners of the internet into classrooms, group chats, and, increasingly, courtrooms.
The Science Behind the Scandal
The technology itself is not mysterious. Diffusion and generative adversarial models are trained on enormous image datasets and can now convincingly reconstruct what a clothed body “might” look like unclothed, blending a real face onto a fabricated one. What once required expensive software and technical skill is now a free app or a few dollars on a website, accessible to any teenager with a smartphone. The barrier to creating exploitative imagery has collapsed even as the emotional and legal consequences have not.
The scale is staggering. According to the National Center for Missing and Exploited Children, reports of AI-generated child sexual abuse material submitted to its CyberTipline jumped from roughly 4,700 in 2023 to more than 440,000 in just the first half of 2025. The FBI’s Internet Crime Complaint Center has been blunt about the legal reality: content manipulated or generated by AI to depict a minor in a sexually explicit manner is illegal, regardless of whether the underlying photo was innocent or the child in the final image is a composite.
Classrooms Turned Crime Scenes
The human cost has already played out publicly. At Lancaster Country Day School in Pennsylvania, two 14-year-old boys used AI tools to generate roughly 350 fabricated nude images of at least 59 female classmates, pulling source photos from yearbooks, social media, and FaceTime calls. Both pleaded guilty to dozens of felony counts of sexual abuse of children and were sentenced to juvenile probation, community service, and restitution a sentence that left victims’ families arguing the punishment did not match the trauma inflicted.
New Jersey had its own reckoning in October 2024, when boys at a Westfield high school circulated AI-generated nude images of female classmates. The case became a statewide flashpoint after student Francesca Mani spoke out publicly, pressing lawmakers to close the gap between rapidly evolving technology and outdated statutes.
The Law Catches Up Slowly
New Jersey responded faster than most. In April 2025, Governor Phil Murphy signed P.L. 2025, c.40, making New Jersey among the first states to explicitly criminalize the creation and distribution of AI-generated sexual deepfakes, including images of minors. The law establishes it as a third-degree crime, carrying potential imprisonment and fines up to $30,000, while also creating a civil cause of action allowing victims to recover up to $1,000 per fabricated image.
For families, this creates a genuinely difficult legal landscape. A teenager who thought he was making a “joke” image in a group chat may now be facing charges that overlap with existing child exploitation and endangerment statutes, the same statutory framework historically reserved for CSAM produced by adults. Because these cases sit at the intersection of juvenile court procedure, sex offense classification, and now brand-new AI-specific statutes, families navigating an arrest or investigation benefit from guidance that understands all three layers at once. Securing legal counsel in New Jersey early, often before formal charges, is critical in determining whether a case is resolved through diversion or follows a far more damaging adult criminal path.
The Deeper Question Beneath the Code
Strip away the legal statutes and the machine learning architecture, and what remains is an old question dressed in new technology: what do we owe another person’s image, their dignity, their control over how they are seen? Faith traditions have wrestled for millennia with the ethics of representation and consent. Adolescent psychology tells us that teenage brains are still developing the impulse control and empathy needed to grasp the permanence of a digital act. Combine an underdeveloped prefrontal cortex with a frictionless tool that can inflict harm in seconds, and the result is unsurprising; it is nearly predictable.
What Schools and Families Can Do Now
Prevention starts with treating these apps the way communities eventually treated cyberbullying: not as a prank, but as a documented harm with real victims. Schools are beginning to build explicit AI-image policies into their codes of conduct. Parents are having earlier, more specific conversations about consent and digital image manipulation, not just “stranger danger.” And when a case does surface, speed matters: evidence preservation, how school officials or police question a student, and whether a case is diverted or prosecuted can all hinge on decisions made in the first 48 hours.
The technology will keep evolving faster than any single law can anticipate. What communities can control is how seriously they treat the moral weight of an image the moment it exists, long before a court ever has to decide what it means.
