ChatGPT Was Reportedly Used in Planning School Stabbing in Pirkkala, Finland, That Injured Three Pupils

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.

Risk classification

• Primary risk domain: 1 Discrimination & Toxicity
• Primary risk subdomain: 1.2 Exposure to toxic content

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.

Additional risk subdomains

• 7.3 Lack of capability or robustness: The AI system demonstrated a lack of safety robustness by failing to prevent the generation of instructions for planning a violent physical attack.

Causal factors

• Entity: Human
• Intent: Intentional
• Timing: Post-deployment

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.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

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.

AI system and alleged parties

• AI system: ChatGPT (OpenAI)
• AI purpose: Chatbot; Question Answering
• Behaviour type: Assistant
• Alleged developer: OpenAI, AI search technology developers
• Alleged deployer: Pirkkala school stabbing suspect, OpenAI
• Alleged harmed parties: Vähäjärvi School pupils, Vähäjärvi School community, Pirkkala school stabbing victims

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Negligible. Severity is scored from Negligible to Catastrophic in each harm category, for harm the reports describe as caused directly or indirectly by the AI system.

• Physical: direct Negligible, indirect Minor
• Infrastructure: direct Negligible, indirect Negligible
• Property: direct Negligible, indirect Negligible
• Financial: direct Negligible, indirect Negligible
• Environmental: direct Negligible, indirect Negligible
• Malicious content: direct Negligible, indirect Minor
• Differential treatment: direct Negligible, indirect Minor
• Civil rights: direct Negligible, indirect Minor
• Democracy: direct Negligible, indirect Negligible
• Privacy: direct Negligible, indirect Negligible
• Psychological: direct Negligible, indirect Minor
• Epistemic: direct Negligible, indirect Negligible
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Physical

Reported: The report explicitly describes physical injuries to three female pupils who were stabbed during the attack.

Indirectly caused: Three female pupils sustained non-life-threatening injuries from a bladed weapon, with one stabbed in the neck, one in the hand, and one in the waist.

Malicious content

Reported: The report explicitly describes an online manifesto published by the suspect detailing his violent intentions.

Indirectly caused: The suspect published a manifesto online detailing his plans to stab girls and claiming he used ChatGPT to prepare for the attack.

Differential treatment

Reported: The report explicitly describes that the suspect deliberately targeted female pupils.

Indirectly caused: The attacker targeted female pupils, stabbing three girls based on gender-targeted animus described in his manifesto.

Civil rights

Reported: The report explicitly describes a violent attack that violated the victims' fundamental right to life and physical security.

Indirectly caused: The physical attack violated the fundamental human rights to bodily integrity and personal security of the three female pupils who were stabbed.

Psychological

Reported: The report explicitly describes confusion, fear, and shock within the school and local community during and after the incident.

Indirectly caused: Students experienced fear and confusion while confined to classrooms during the lockdown, and the local community was deeply shocked by the unprecedented violence.

Inferred additional harm: It is highly likely that the three victims, the attacker, and many of the 1,200 students and staff members present during the lockdown suffered acute psychological trauma and distress requiring ongoing support.

People affected

• Occurrences reported: 1
• People reportedly harmed: 3
• People reportedly exposed: 1200

Potential causes

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.

Information quality

• Classification confidence: Medium
• Reason for confidence: 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.
• Ambiguities identified: The authorship of the manifesto and the actual extent of ChatGPT's involvement in the planning phase remain unconfirmed by official police sources in the provided text.
• Alternative interpretations: 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.
• Missing information: 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.

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.

• Overall national security impact: Minor
• Scope: Single nation
• Primary target: Finland
• Alleged perpetrator: 16-year-old boy

Threat characteristics

• Imminence: Long-term. The immediate physical threat has been neutralized with the suspect in custody. The ongoing concern is the long-term policy challenge of preventing AI misuse for planning violence.
• Autonomy: Human-controlled. The AI acted strictly as a passive informational assistant, with all planning decisions, targeting, and physical execution carried out entirely by the human suspect.
• Novelty: Evolved capability. While using digital tools to plan violence is established, leveraging large language models to systematically research attack methods and evidence concealment represents an evolving threat vector.

Impact by dimension

• Physical security: Minor. Localized violent attack resulting in physical injuries to three individuals, resolved by local law enforcement with no threat to critical infrastructure or national defense systems.
• Information security: Negligible. No indication of information warfare, intelligence compromise, state-sponsored propaganda, or classified data theft.
• Sovereignty: Negligible. Core government functions, state authority, and constitutional processes remained entirely unaffected by this local criminal incident.
• Economic security: Negligible. No impact on national economic stability, strategic industries, intellectual property theft, or technological competitive advantage.
• Societal stability: Minor. The incident caused localized panic and trauma within a school community of 1,200 students, but did not threaten large-scale societal stability or systemic human rights.