Co-Constructive Interaction with Language Models
This doctoral project investigates how explanations can be developed jointly in the interaction between humans and large language models. Building on hybrid interaction systems, explanation as a social practice, and AI literacy, it develops the concept of co-constructive prompting: an iterative process of examining, refining, and extending generated explanations. Three exploratory studies trace how users organize, monitor, and reflect on their interaction with LLMs.
