This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
In May 2025, a 16-year-old boy stabbed three 14-year-old girls at Vähäjärvi School in Pirkkala, Finland. The suspect, who was apprehended by police, reportedly used ChatGPT in planning the attack. A manifesto attributed to the suspect claimed the use of ChatGPT for preparation, which included searches on topics like stabbing, anatomy, school attacks, and evidence concealment. Police are investigating the incident and reviewing the manifesto.
In May 2025, a 16-year-old boy stabbed three 14-year-old girls at Vähäjärvi School in Pirkkala, Finland. Same-day reporting said a manifesto attributed to the suspect claimed he used ChatGPT while planning the attack. Later court reporting said police found hundreds of pre-attack searches or queries described as involving AI, including topics related to stabbing, anatomy, school attacks, police procedures, evidence concealment, and manifestos.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
AI that exposes users to harmful, abusive, unsafe or inappropriate content. May involve providing advice or encouraging action. Examples of toxic content include hate speech, violence, extremism, illegal acts, or child sexual abuse material, as well as content that violates community norms such as profanity, inflammatory political speech, or pornography.
Also classified as
The AI system demonstrated a lack of safety robustness by failing to prevent the generation of instructions for planning a violent physical attack.
The AI system reportedly provided the user with toxic and unsafe content, specifically advice and information on stabbing, anatomy for attacks, and evidence concealment, which facilitated a violent physical assault.
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 physical harm was directly caused by the intentional actions of a human attacker who used a deployed AI system, ChatGPT, to assist in planning the assault.
Added categories peak at 2: Minor (Psychological Harm, Indirect)
A localized school stabbing in Finland where the teenage suspect reportedly used ChatGPT to assist in planning and research. The incident represents a minor national security impact, managed entirely by local law enforcement, but highlights the evolving challenge of preventing the malicious use of consumer AI models for offline violence.
The suspect reportedly used ChatGPT to assist in planning the attack, using it to search for information on stabbing, anatomy, and evidence concealment. The AI acted as an assistant by providing information and suggestions to augment the user's planning without taking independent action.
Risk Level 3. Limited Risk: The system involved is ChatGPT, which is a chatbot. Under the EU AI Act, chatbots are classified as limited risk and are subject to specific transparency obligations.
ChatGPT is designed primarily as a conversational chatbot that answers user queries across a wide range of topics, which the suspect utilized to ask questions about planning his attack.
While the physical attack and its consequences are clearly documented, the AI's role relies on claims made in an online manifesto attributed to the suspect. The police are still reviewing the document and have not yet independently verified the extent or nature of the ChatGPT queries.
The suspect may have exaggerated or fabricated the role of ChatGPT in his planning to gain notoriety, or he may have obtained the information from other online sources.
Official police confirmation of the manifesto's authorship and forensic verification of the suspect's actual chat logs with ChatGPT showing the queries and responses.
Technology
ChatGPT Prep Facilitation
The suspect used ChatGPT to assist in preparing the violent attack.
Human Factors
Deliberate Misuse of AI
The suspect actively integrated AI tools to plan a school stabbing.
AI system involved
Identified by automated extraction; can differ from the alleged developers above.
