This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
Schools across Australia are experiencing a surge in the use of AI-powered 'nudify' apps to create and circulate non-consensual deepfake pornography of students and teachers. Perpetrators, often students themselves, use these tools to generate graphic images from social media photos, leading to severe psychological trauma for victims. The incident at Bacchus Marsh Grammar, where a student was arrested for creating images of 50 classmates, highlights the widespread nature of this abuse. Regulators and the government are now working to implement stricter laws and standards to hold platforms and users accountable for this form of image-based abuse.
Throughout 2024, schools in Australia dealt with a significant rise and proliferation of non-consensual deepfake pornography of students. Often, male students are reported to use "nudify" apps such as Undress AI with images of their classmates and teachers. Many of the sites have remained legal and accessible to minors, who in turn are using the sites to generate pornography of their peers.
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 incident involved the generation and circulation of highly graphic, non-consensual sexual imagery depicting minors (CSAM).
The perpetrator bypassed private social media settings to harvest photos and generate intimate, non-consensual depictions of the victims.
The incident involves the targeted creation of humiliating and sexual deepfake imagery of specific victims for personal harassment 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 harm was driven by the intentional actions of human users utilizing a deployed AI tool to target and humiliate specific individuals.
Added categories peak at 5: Catastrophic (Child Sexual Exploitation & Abuse (CSEA), Inferred)
The incident involves the widespread use of AI-powered 'nudify' apps by students to generate non-consensual explicit imagery of classmates and teachers in Australia. While causing severe psychological trauma and prompting legislative reforms, it remains a domestic law enforcement and regulatory issue rather than a threat to national security.
The AI system, Undress AI, operates as a tool. It performs a specific, well-defined function (generating deepnude images) based on direct user inputs (uploading a photo, selecting settings like age and body type) and produces deterministic outputs in less than 30 seconds without maintaining an independent goal or planning capabilities.
Limited Risk: The system is used to generate AI-generated content such as deepfakes, which falls under Limited Risk and carries specific transparency obligations under the EU AI Act.
The primary function of the Undress AI system is to generate realistic deepfake nude images of individuals by manipulating uploaded photos.
The reports provide clear, detailed accounts of the specific incidents at Bacchus Marsh Grammar and Salesian College, including the number of victims, the specific AI tool used, and the psychological impacts. The role of the AI system is explicitly stated.
None
Technology
Rapid Generation Speed
AI generates realistic deepnudes in less than 30 seconds.
Low Technical Barrier
No advanced skills or computing power needed compared to Photoshop.
Lack of Safeguards
Undress AI apps lack controls to block generating child images.
Data Inputs
Social Media Harvesting
Photos are easily harvested from public or private social media accounts.
Scraped Training Data
AI models are trained on real images scraped from across the internet.
Human Factors
Malicious User Intent
Students and predators intentionally creating non-consensual deepnudes.
Lack of Ethical Awareness
Schoolboys generating explicit images of classmates without understanding harm.
Process and Methods
Easy Online Accessibility
Deepnude apps are readily available via simple Google searches.
Social Media Distribution
Generated images are easily shared via Instagram and Snapchat.
Ineffective Removal Processes
Difficulties in quickly removing distributed explicit content online.
Management
Negligent App Operators
Parent companies fail to implement age or consent verification.
Pipeline Accountability Failure
Search engines and payment processors facilitate deepnude distribution.
Lack of Risk Assessment
AI developers fail to assess risks of child exploitation before launch.
Regulatory Environment
Gaps in Existing Laws
Federal laws did not adequately cover adult deepfake sexual material.
Lagging Legal Frameworks
Legislation is catching up slowly to rapid AI technological advances.
Lack of Deplatforming Powers
Regulators lacked immediate standards to take offending apps offline.
AI system involved
Identified by automated extraction; can differ from the alleged developers above.
