Associations between generative AI use frequency, technology acceptance, attitudes toward AI, and reported learning...

Physical education combines embodied practice with cognitive learning, yet generative artificial intelligence is increasingly used as an academic support tool. This cross-sectional study investigated associations between self-reported generative AI use frequency, reported VARK learning preference patterns, attitudes toward AI, and Technology Acceptance Model beliefs among undergraduate students in PE classes.

A cross-sectional survey was conducted with 1,084 full-time PE undergraduates (age 19-28 years). Participants completed PE-adapted instruments assessing reported learning preferences patterns using the VARK framework, technology acceptance beliefs based on the Technology Acceptance Model, attitudes toward generative AI, and self-reported generative AI use frequency. The TAM and VARK instruments were adapted to PE-relevant academic and movement-related tasks. Data were analyzed using descriptive statistics, reliability testing, ANOVA, and multiple regression models.

Students reported moderate acceptance of generative AI and positive attitudes toward its academic use. Kinesthetic learning remained the dominant preference, consistent with the movement-based nature of PE. However, higher self-reported AI usage frequency was associated with higher visual and reading-writing scores and lower kinesthetic scores (β = .34-.40; β = -.39, p < .001). Significant differences emerged across learning-style groups for all TAM and AI attitude dimensions (p < .001, η ² = .086-.142). Regression analysis showed that perceived usefulness, ease of use, and positive attitudes toward AI were the strongest predictors of AI usage, explaining 48% of variance. Frequent AI engagement was associated with greater concentration in in reported VARK score patterns (β = -.62, p < .001).

In this cross-sectional sample, self-reported generative AI use was associated with differences in reported learning preference patterns and with stronger technology acceptance beliefs among higher education physical education students. While kinesthetic scores remained highest overall, more frequent AI use was associated with higher visual and reading-writing scores and lower kinesthetic scores. These findings are associative, not causal, and do not show changes in stable learning preference patterns. Future research should examine whether pedagogically grounded AI integration can support diverse learners, particularly students who report stronger kinesthetic preferences.