Agentic Engagement with Educational AI Chatbots Among Pre-service Teachers: A Mixed-Method Study in Qatar

Abstract

Agentic engagement captures students’ intentional efforts to shape their learning environment and co-create their educational experiences. However, there is limited understanding of why engagement with educational artificial intelligence (AI) chatbots varies among students, particularly among pre-service teachers, and agentic engagement is often overlooked in AI-mediated instruction. This mixed-method study explored engagement with an AI chatbot among 33 pre-service teachers enrolled in a classroom management course at one university in Qatar. Guided by self-determination theory and agentic engagement, we analysed 2,189 chatbot messages and end-of-course interviews, and classified participants into high, medium, and low engagement groups. The findings showed that high-engagement students used the AI chatbot autonomously for clarification and assignments, although they expressed minimal ethical concerns. Medium-engagement students balanced selective use with collaboration and raised concerns about plagiarism. Low-engagement students engaged only sporadically, placed greater value on instructor guidance, and expressed mixed ethical views. Overall, the results highlighted the potential of AI chatbots to support autonomy and competence, while also revealing gaps in relatedness and risks of overreliance. This study contributes to engagement theory by showing that frequency of use can mask different forms of agentic engagement, including selective use and agentic refrain. It also underscores the need for thoughtful integration, supported by instructor guidance, to promote balanced, ethical and reflective engagement.