This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
In December 2023, female students at Collège Béliveau in Winnipeg were targeted by the creation and online distribution of non-consensual AI-generated nude images. The perpetrators reportedly used publicly accessible social media photos to generate the explicit content. The incident caused significant emotional distress among the student body and prompted an investigation by the Winnipeg Police Service, though no criminal charges were ultimately filed due to legal complexities surrounding AI-generated content.
At Collège Béliveau in Winnipeg, female students between grades 7-12 were targeted in the creation of deepfake nudes, which were then distributed online. Specific numbers and identities of victims and perpetrators were not released, and no charges were ultimately filed owing to the gap between existing laws and the nature of the incident.
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 AI was used to generate and distribute child sexual abuse material (CSAM), exposing the school community to highly toxic and illegal content.
The incident involves the targeted creation of humiliating and sexualized deepfake imagery of specific underage victims for personal amusement or malicious intent.
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 caused by human actors intentionally using deployed AI deepfake tools to alter and distribute images of students.
Added categories peak at 3: Substantial (Child Sexual Exploitation & Abuse (CSEA), Direct)
An incident involving the creation and distribution of AI-generated non-consensual explicit images of teenage students in Winnipeg, Canada. While causing significant localized psychological harm and highlighting legal gaps regarding synthetic CSAM, the event has negligible direct implications for national security.
The AI systems were used as simple tools by human actors to perform a specific, well-defined function (generating deepfakes from input photos) without autonomous planning or context retention.
Limited Risk: The report describes the use of AI-generated content such as deepfakes, which falls under Limited Risk and carries specific transparency obligations.
The AI tools were explicitly used to take existing photos of students and alter them to make them appear nude, which is the definition of deepfake image generation.
The reports provide consistent, detailed accounts of the incident at Collège Béliveau, including the scale of the images (upwards of 300), the demographic targeted (dozens of female students), and the legal/educational responses. While the specific AI software is not named, the nature of the technology and its application are clearly described.
None. The incident is clearly a case of non-consensual deepfake creation targeting minors.
Technology
Highly Accessible Deepfake Tools
AI tools are increasingly advanced, free, and easy for anyone to use.
Realistic Image Manipulation
AI allows seamless face-swapping and clothing removal with high realism.
Data Inputs
Public Social Media Photos
Original student photos were gathered from publicly accessible social media.
Human Factors
Ignorance of Legal Consequences
Perpetrators do not understand that deepfakes are child abuse material.
Lack of Digital Literacy
Students are ill-equipped to handle online risks or know their rights.
Process and Methods
Outdated School Curriculum
Curriculum fails to address online sexual violence and legal consequences.
Difficult Image Removal Process
Removing images permanently from the internet is nearly impossible.
Management
Profit-Driven AI Development
Tech firms rush out AI for profit without ethics or safety guardrails.
Lack of Proactive Risk Assessment
Failure to assess risks of image generators before releasing them.
Regulatory Environment
Legal Gaps for AI Deepfakes
Many intimate image laws were created before deepfakes and lack key terms.
Extraterritorial Jurisdiction
Enforcing Canadian court orders on foreign websites is highly challenging.
Section 230 Liability Shield
US sites can ignore Canadian removal orders due to Section 230 protections.
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.
