Answers are grounded in this paper and cite their sources.
This study investigated how teachers interacted with a Large Language Model (LLM)-based chatbot and the pragmatic strategies they employed during these interactions. Three elementary school teachers used the chatbot to learn a pedagogy called knowledge building and to co-design knowledge building lessons. The chatbot's functional responses were shaped by several factors, including its pretraining, the system prompts provided by the researcher, and the additional curated content linked to the LLM. Within this context, the chatbot demonstrated a reasonable level of functional execution and content accuracy related to knowledge building when responding to teachers’ pragmatic moves. The three teachers exhibited distinct interaction patterns—ranging from information seeking, to clarification and questioning, to collaborative idea improvement—which resulted in varying quantity and depth of ideas explored. The findings offer insights into how to optimize interactions with LLM-based chatbots to support teachers' productive learning of knowledge building.
No takes yet. Share an insight, caveat, or question.
Tan et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: