Evidential value of AI chatbots for common general knowledge

The widespread availability and use of chatbots based on large language models or artificial-intelligence techniques does not, by itself, justify assuming that an answer generated by such a system accurately reflects the skilled person’s common general knowledge in the relevant technical field at the relevant date. In particular, the answer is based on training data that are opaque to the user and may depend heavily on the context and precise wording of the prompt (EPO BoA 15 Apr 2025, T 1193/23).