Hao-Chuan Wang
Author directory2026
How Persona and Message Type in Conversation Initiation Influence Learners’ Experience with Educational Chatbots
Jingchao Fang | Qingxiaoyang Zhu | Hao-Chuan Wang
Proceedings of the Fifth Workshop on NLP for Positive Impact (NLP4PI)
Jingchao Fang | Qingxiaoyang Zhu | Hao-Chuan Wang
Proceedings of the Fifth Workshop on NLP for Positive Impact (NLP4PI)
Video-based learning is widely adopted among online learners, but it often lacks interactivity and fails to engage learners. Advances in large language models (LLM) show promise in enabling role-playing characters as learning companion chatbots for interactive learning. While much research has focused on LLM-powered chatbots for question-answering or responding, fewer studies have explored their role as proactive conversation initiators and prompt learners to think and reflect. How do different styles of chatbots’ conversation initiation influence learners’ video-based learning experience? Through controlled empirical studies with learners (N=65), we find that learning companion chatbot’s personas and message styles influence learners’ perceptions, reactions, and overall learning experiences. Our findings provide key insights for designing engaging, effective education chatbots.