Chatbot versus preservice teacher-led reading in primary schools: the roles of engagement and learner differences

Artificial intelligence (AI) chatbots have the potential to enhance learning outcomes by assisting teachers with time and skill limitations. However, their specific roles in reading instruction and the psychological mechanisms involved are not well understood. This study aims to explore whether, why, and how interactive reading, a technique involving reading aloud, dynamic questioning, and feedback, benefits primary school students’ reading comprehension performance when led by AI chatbots. The study involved 127 primary school students from mainland China, who were assigned to one of the three groups for plain text reading: AI chatbot-led, preservice teacher-led, or independent reading. Results showed that AI chatbot-led reading not only improved overall reading comprehension scores comparably to preservice teacher-led reading but also achieved higher inferential reading scores. The facilitative effect of AI chatbot-led reading was moderated by students’ baseline reading abilities and reading motivation. In addition, the coded engagement score based on video annotation mediated the conditional effect on inferential reading scores. These findings indicate that AI chatbots can effectively improve children’s reading performance, particularly for primary school students with lower baseline reading abilities and motivation, by acting as engaging and supportive facilitators. This study highlights the potential of AI chatbots in enhancing reading instruction.