AI Hallucinations: Whoever Operates a Chatbot Is Liable for Its Statements (Higher Regional Court of Hamm, Regional Court of Hamburg)

AI is not a person in its own right. Nobody can "excuse" themselves with it and escape liability with the "argument" that the AI hallucinated. Deployers of AI too (see the definition in Art. 3 no. 4 of the EU AI Act: "deployer" means a natural or legal person, public authority, agency or other body using an AI system under its authority, except where the AI system is used in the course of a personal non-professional activity) adopt the output of the AI as their own. They are liable for it if it is wrong. This is increasingly borne out by court decisions:

Already in Moffatt v. Air Canada (decision of 14 February 2024 – SC-2023-005609) of the Civil Resolution Tribunal (CRT) of British Columbia, Canada, Air Canada was held responsible for the output of its chatbot.

This responsibility has now also arrived in the German courts. The Regional Court (Landgericht, LG) of Hamburg (order of 23 September 2025 – 324 O 461/25) had already decided this in connection with the chatbot Grok. It held the operator X.AI LLC liable for AI-generated statements of the X account Grok.

The Higher Regional Court (Oberlandesgericht, OLG) of Hamm (judgment of 12 May 2026, case no. 4 UKl 3/25) ruled a few days ago: whoever operates an AI chatbot is liable for its false statements.

Conclusion: The use of AI should be well thought through. Otherwise, hallucinations in uncontrolled output become an incalculable risk.