Abstract

Natural language processing has taken enormous steps during the last few years. The development of large language models and generative AI has elevated natural language processing to the level that it can output coherent and contextually relevant text for a given natural language prompt. ChatGPT is one incarnation of these steps, and its use in education is a rather new phenomenon. In this paper, we study students’ perception on ChatGPT during a computer science course. On the course, we integrated ChatGPT into Teams private discussion groups. In addition, all the students had freedom to employ ChatGPT and related technologies to help them in their coursework. The results show that the majority of students had at least tested AI-powered chatbots, and that students are using AI-powered chatbots for multiple tasks, e.g., debugging code, tutoring, and enhancing comprehension. The amount of positive implications of using ChatGPT takes over the negative implications, when the implications were considered from an understanding, learning and creativity perspective. Relatively many students reported reliability issues with the outputs and that the iterations with prompts might be necessary for satisfactory outputs. It is important to try to steer the usage of ChatGPT so that it complements students’ learning processes, but does not replace it.