University students’ AI thinking: A Latent Profile Analysis and predictive effects of psychological factors

Despite the growing presence of artificial intelligence (AI) in education, limited research has examined the psychological profiles underlying students' cognitive engagement with AI. Using Latent Profile Analysis, this study identified four distinct AI thinking profiles among university students: Low-Engagement Thinkers (19.6%), Moderately Engaged Thinkers (23.4%), Engaged and Critical Thinkers (23.9%), and Highly Engaged Thinkers (33.0%). Multinomial logistic regression revealed that negative academic emotions significantly predicted membership in all higher-engagement groups, with the strongest effect observed for Highly Engaged Thinkers. Test anxiety also emerged as a key predictor for this group. These findings highlight the diverse nature of students' AI cognition and emphasize the role of emotional strain in driving engagement. Practical implications for emotionally aware AI integration in education are discussed.