This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
In August 2025, students at a Louisiana middle school used generative AI to create and circulate sexually explicit deepfake images of classmates and adults. The incident resulted in significant psychological harm to the victims and led to a physical altercation on a school bus, after which one of the victims was expelled. The case highlights the growing challenge of AI-enabled cyberbullying in schools and the inadequacy of existing institutional responses to such incidents.
Purportedly AI-generated sexually explicit images depicting a 13-year-old student and other middle school students were reportedly created and circulated at a Louisiana school. The images were shared via social media and shown to students on campus, resulting in ongoing harassment and psychological harm. After the images were displayed on a school bus, the affected student became involved in a physical altercation and was expelled.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
Using AI systems to gain a personal advantage over others such as through cheating, fraud, scams, blackmail or targeted manipulation of beliefs or behavior. Examples include AI-facilitated plagiarism for research or education, impersonating a trusted or fake individual for illegitimate financial benefit, or creating humiliating or sexual imagery.
Also classified as
The creation and viral spread of explicit deepfakes exposed minors and the school community to highly toxic and inappropriate content.
The incident involved the creation of humiliating and sexually explicit synthetic imagery of specific middle school students to harass and humiliate them, which falls under targeted manipulation and abuse.
Human
Due to a decision or action made by humans
Intentional
Due to an expected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed
The incident was caused by students intentionally using post-deployment generative AI tools to create and distribute non-consensual explicit images of their classmates.
Added categories peak at 3: Substantial (Child Sexual Exploitation & Abuse (CSEA), Direct)
A localized incident involving Louisiana middle school students using generative AI to create non-consensual explicit images of classmates. While causing significant local distress and highlighting gaps in educational policy, the incident has negligible national security implications and was managed via local law enforcement and school disciplinary actions.
The report describes the technology as an app or social media tool that allows users to 'pluck a photo off social media, nudify it' and generate a deepfake. This represents a Tool behavior where the AI performs a specific, well-defined function on direct input.
Risk Level 3. Limited Risk: The AI system is used to generate deepfakes. The EU AI Act classifies 'AI-generated content such as deepfakes' under Limited Risk, which carries specific transparency obligations.
The AI system was used specifically to generate realistic, non-consensual nude deepfake images of individuals by manipulating existing photos.
The reports provide detailed, consistent accounts of the incident, including the timeline, the number of victims (8 students, 2 adults), the school's response, and the legal outcomes. The role of generative AI in creating the deepfakes is explicitly stated.
None
Technology
Easy-to-use AI Nudification Apps
AI tools allow creating realistic nudes with no technical expertise.
Ephemeral Messaging Platforms
Snapchat deletes images quickly, making AI deepfakes hard to trace.
Realistic Image Generation
AI generates highly realistic fakes that are difficult to prove false.
Data Inputs
Social Media Photos
Publicly accessible photos of minors were used as inputs for AI tools.
Human Factors
Adult Skepticism and Disbelief
School administrators dismissed reports of AI deepfakes as teenage rumors.
Lack of Student Digital Empathy
Students shared AI deepfakes without understanding the severe trauma caused.
Victim Emotional Distress
Unrelenting harassment from AI nudes led to the victim's physical outbreak.
Process and Methods
Outdated Cyberbullying Curriculum
School's safety curriculum from 2018 lacked guidance on AI deepfakes.
Inadequate Verification Processes
No protocol to quickly verify AI deepfakes on ephemeral applications.
One-Sided Disciplinary Actions
Victim was expelled immediately while deepfake creators faced delayed action.
Management
Ostrich Syndrome on AI Risks
School management avoided proactively addressing risks of AI-generated nudes.
Inadequate Risk Assessment
Management failed to assess how generative AI could weaponize student photos.
Regulatory Environment
Lagging School AI Policies
District policies only addressed academic AI use, ignoring cyberbullying.
Delayed Legal Accountability
New state laws on AI deepfakes exist but enforcement is often retroactive.
AI system involved
An AI system was involved, but the reports do not identify it.
Identified by automated extraction; can differ from the alleged developers above.
