Tunisian Man Allegedly Used ChatGPT to Research Bomb-Making During Paris Terror Plot Targeting Louvre and Jewish Community

French investigators alleged that Dhafer Mellala used ChatGPT to research the construction and destructive effects of explosive devices while planning a jihadist-inspired attack in Paris. His communications reportedly discussed placing explosives at the Louvre and attacking members of the Jewish community in the 16th arrondissement. He denied planning an attack, saying his activity reflected curiosity.

Risk classification

• Primary risk domain: 4 Malicious actors
• Primary risk subdomain: 4.2 Cyberattacks, weapon development or use, and mass harm

The suspect used ChatGPT to research weapon development, specifically how to manufacture a bomb and obtain chemical precursors for an explosive attack.

Causal factors

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

The risk was driven by a human actor intentionally using a post-deployment AI system to facilitate a terror plot.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Limited Risk: ChatGPT is a general-purpose conversational chatbot. According to the EU AI Act definitions, chatbots pose a moderate risk and are subject to transparency obligations to ensure users are aware they are interacting with an AI.

AI system and alleged parties

• AI system: ChatGPT (OpenAI)
• AI purpose: Chatbot; Question Answering
• Behaviour type: Assistant
• Alleged developer: OpenAI, Large language model developers, Chatbot developers
• Alleged deployer: Terrorism suspects, Malicious actors, Dhafer Mellala
• Alleged harmed parties: National security and intelligence stakeholders, Musée du Louvre visitors and staff, Musée du Louvre, Jewish people, Jewish community of France, General public of France, General public

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 Negligible
• 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 Negligible
• Differential treatment: direct Negligible, indirect Negligible
• Civil rights: direct Negligible, indirect Negligible
• Democracy: direct Negligible, indirect Negligible
• Privacy: direct Negligible, indirect Negligible
• Psychological: direct Negligible, indirect Negligible
• Epistemic: direct Negligible, indirect Negligible
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

People affected

• Occurrences reported: 1

Potential causes

Technology

• Delayed Refusal: ChatGPT processed queries on TNT damage before eventually refusing.
• Failure to Block Bomb Queries: AI allowed searches on bomb-making and chemical precursors.

Human Factors

• Malicious AI Exploitation: Suspect used ChatGPT to find bomb-making info and precursor data.

Information quality

• Classification confidence: High
• Reason for confidence: The reports clearly identify the suspect, his arrest, the specific AI model used (ChatGPT), and the nature of his queries ('how to make a bomb', TNT damage). There is minimal ambiguity regarding the AI's role as a tool queried by the suspect, and the reports explicitly note that ChatGPT refused to answer at least one dangerous query.
• Ambiguities identified: It is not fully clear whether ChatGPT successfully answered the initial 'how to make a bomb' query or if it blocked all bomb-related queries, though it is explicitly stated that it refused the TNT damage query.
• Alternative interpretations: None. The incident is a clear case of a malicious actor attempting to leverage a public LLM for weapon development.

A Tunisian national was arrested in Paris for plotting a terror attack targeting the Louvre and the Jewish community. The suspect utilized ChatGPT to research bomb-making and chemical precursors. The plot was disrupted by law enforcement before execution, resulting in minor national security impact but highlighting the ongoing risk of malicious actors attempting to leverage LLMs for weapon development.

• Overall national security impact: Minor
• Scope: Single nation
• Primary target: France
• Alleged perpetrator: Dhafer M.

Threat characteristics

• Imminence: Long-term. The immediate threat was neutralized by the suspect's arrest, leaving an ongoing strategic concern regarding terrorist exploitation of LLMs.
• Autonomy: Human-controlled. The AI acted solely as a query-response assistant, with all planning, intent, and decision-making driven entirely by the human user.
• Novelty: Evolved capability. Represents an incremental advancement where malicious actors attempt to exploit public generative AI models rather than traditional search tools for weapon research.

Impact by dimension

• Physical security: Minor. The terror plot targeting the Louvre was disrupted by law enforcement before execution. The AI was used for informational research on explosives, presenting a minor physical security threat.
• Information security: Negligible. No intelligence compromise, classified data theft, or information warfare operations were indicated in the incident.
• Sovereignty: Negligible. No direct threat to state authority, elections, or core government decision-making processes was identified.
• Economic security: Negligible. No significant threat to economic stability, financial systems, or strategic technological competitive advantage.
• Societal stability: Minor. The plot targeted a specific religious community, which could have threatened social cohesion, but it was successfully disrupted before causing harm.