Abstract
ABSTRACT
This study investigated the relationships between learner characteristics, AI utilization, and learning outcomes in an intelligent nutrition education platform. By examining 109 university students, we explored how prior nutrition literacy, motivational factors, task anxiety, and critical thinking abilities influenced learners' interactions with AI support and their subsequent learning experiences. The study employed both quantitative analysis of pre‐test and post‐test measures, as well as qualitative content analysis of AI interactions. Results revealed that participants with higher baseline nutrition literacy were less likely to utilize AI support, demonstrating an expertise reversal effect. Although prior nutrition literacy significantly influenced feelings of knowing (FOK), AI usage showed no significant direct impact on perceived usefulness or judgment of learning. Analysis of AI interactions revealed a strong preference for factual information seeking, with limited engagement in normative and interpretative queries. These findings suggest the need for adaptive AI systems that can calibrate support based on prior knowledge and highlight the importance of scaffolding more sophisticated forms of AI interaction in educational settings. This research contributes to our understanding of how AI tools mediate learning processes and informs the design of more effective smart learning environments for nutrition education.