AI deepfake nudes become a new school bullying crisis

The Wall Street Journal has reported a disturbing shift in how generative AI is being misused by children and teenagers: so-called nudify tools are turning ordinary photos into fake explicit images and creating a new form of bullying inside schools. The article, published on June 13, 2026, describes how parents, school leaders and victims are struggling to respond as the tools become cheap, easy to find and simple enough for students to use.

The core problem is not only that the images are fake. It is that they are believable enough to damage a student's reputation, spread quickly through group chats and social platforms, and leave victims feeling powerless even after adults confirm that the image was generated. For schools, that creates a hard operational problem. Traditional bullying policies were built around insults, threats, fights or real photos. AI-generated sexual images sit across discipline, child protection, technology policy, law enforcement and mental health support.

The WSJ report also shows why platform accountability is becoming central. The apps and websites that enable this abuse often sit outside the school environment, while distribution happens through messaging services that adults cannot easily monitor. Even when families know what happened, removal can be slow and inconsistent. New U.S. rules such as the Take It Down Act are meant to force faster removal of nonconsensual intimate deepfakes, but the practical burden still often starts with victims, parents and school staff.

For AI users and makers, the lesson is blunt: powerful image generation is not just a creative feature. It can become abuse infrastructure when identity, consent and age protections are weak. Builders of AI image tools, app stores, social platforms and schools all need clearer detection, reporting and escalation paths.

For companies and policymakers, this is another sign that AI safety is moving from abstract model behavior to real-world harm. The next phase of AI governance will not only ask whether a model can generate realistic images, but whether the surrounding product makes misuse harder, removal faster and victims safer.