Different forms of interaction with generative AI in higher education: a latent profile analysis

Generative artificial intelligence (AI) is rapidly transforming how students interact with digital systems in learning contexts, although students differ substantially in how and why they use these technologies. Grounded in Self-Determination Theory and adopting a person-centered approach, this study examined latent profiles of students’ interaction with generative AI based on autonomous motivation, amotivation, AI self-efficacy, and academic self-concept, and their associations with behavioral and attitudinal variables. A sample of 422 undergraduate students completed the study measures. Latent profile analysis identified four distinct user profiles: moderate functional users, self-determined and competent users, low-motivation users, and active competent users. Multinomial logistic regression indicated that AI user identity consistently differentiated the more adaptive profiles, whereas perceiving AI as a form of academic plagiarism was associated with the low-motivation profile. Positive expectations regarding the professional impact of AI were also associated with the self-determined and competent profile. Overall, the findings suggest that students’ relationships with generative AI reflect heterogeneous configurations of motivational regulation, perceived competence, and academic self-perceptions, associated with differences in AI-related identity, ethical beliefs, and professional expectations. These results highlight the importance of motivational heterogeneity in students’ use of generative AI.