Development of cognitive engagement and motivation using AI chatbot-facilitated questioning in medical education

Artificial intelligence tools such as ChatGPT offer new opportunities to support medical students' learning through interactive questioning and instantaneous feedback. However, while ChatGPT may facilitate learning engagement through its accessibility, ChatGPT's potential to foster higher-order thinking and sustained engagement remains underexplored. This study explored how ChatGPT-facilitated questioning relates to medical students' cognitive engagement and motivation in a first-year foundational science module, integrating Self-Determination Theory, Expectancy-Value Theory, and Bloom's Taxonomy as learning frameworks. A study was conducted among medical students pursuing a cardiovascular physiology course. Students generated course-related questions and used ChatGPT to obtain answers, which were subsequently evaluated by instructors. 31 student questions were independently coded by two reviewers according to Bloom's cognitive domains. Perceived autonomy, competence, relatedness, interest/enjoyment, and task value were collected using validated Self-Determination and Expectancy-Value Theory-based tools. Student-generated questions were coded according to Bloom's cognitive domains, and open-ended feedback on the strengths and limitations of ChatGPT was synthesised through conventional content analysis. Survey results indicated higher perceived autonomy, competence, and task value among ChatGPT users compared with non-users, although differences were not statistically significant. Most student-generated questions also reflected lower to intermediate cognitive levels - 'Understand' (41.9%), and 'Apply' (45.2%), with few reaching 'Analyse'. Qualitative insights highlighted ChatGPT's efficiency, accessibility, and cognitive support, alongside concerns regarding accuracy, superficial engagement, and limited interpersonal interactions. Integrated results suggest that ChatGPT may support self-directed motivation and learning but does not consistently facilitate higher-order thinking and social relatedness without instructor mediation. ChatGPT may offer benefits for medical education by supporting autonomy and perceived usefulness. However, motivation alone is insufficient to promote higher-order thinking. ChatGPT should be facilitated by educators to transform artificial intelligence use from information retrieval into reflective, dialogic inquiry. Integrating ChatGPT within collaborative learning may strengthen analytical reasoning and relational engagement in early medical training.