Die Person ChatGPT als „kooperativer Writing-Ko-Aktant“ – über das Spannungsverhältnis von sozialer Rolle und Person im Mensch-Maschine-Dialog

Abstract

This article presents the preliminary findings of a study examining the use of the ChatGPT-4o-mini version of OpenAI's large language model (LLM) for revising student summaries. It should be noted that the statements and conclusions in this study refer exclusively to this LLM in this specific iteration. This study is based on the principles of sociological role theory and subsequent critiques of the theory. To this end, we propose a theoretical framework to delineate the unique requirements of written dialogue between students and the LLM behind ChatGPT. This framework is designed to facilitate coordination between human and artificial intelligence in the text writing process. Following the theoretical considerations, the expectations and assumptions regarding the potential of ChatGPT are examined, as these influence the linguistic and communicative behaviour of LLM users.