Abstract
This report summarizes a focus group held at EuroPLoP 2024 on using ChatGPT as a recommender for design patterns. Apart from the organizers, 11 participants from academia and industry took part in the focus group. In this focus group, we discussed the results of two experiments exploring how well ChatGPT gives recommendations on using patterns from the GoF collection when a situation is described which suggests the application of such a pattern. It was found that the responses are promising in many cases, but their quality depends on a good choice of prompts. The quality of the responses was comparable for prompts in English, German, and Russian, but a little lower for Arabic and absolutely useless in Kyrgyz. Based on our observations, some recommendations are given for selecting suitable prompts when communicating with the chatbot.