The psychological mechanism and heterogeneity of art design students' intention to use AI: an empirical analysis based on an integrated framework.

The rapid development of Artificial Intelligence (AI) is profoundly reshaping the creative industries and art design education. Understanding the factors influencing students' acceptance of Artificial Intelligence generated content is crucial for ensuring the sustainable application of this technology. This study investigated 630 Chinese university students majoring in art and design, adopting a combined variable-centered and person-centered approach to systematically examine the psychological mechanisms and heterogeneity of their intentions to use AI. Based on a serial mediation model, we hypothesized that general attitude toward AI itself and attitude toward using AI serve as mediators between distal antecedents (perceived usefulness, perceived ease of use, AI anxiety, creative self-efficacy) and AI usage intention. At the variable-centered level, hierarchical regression analysis and serial mediation analysis were used to test the predictive and indirect effects of perceived usefulness, perceived ease of use, AI anxiety, creative self-efficacy, general attitudes toward AI itself, and attitudes toward using AI on AI usage intention. At the person-centered level, latent profile analysis was employed to identify heterogeneous user groups based on AI usage intention across five design stages: creative ideation, material collection, visual element design, copywriting, and final modification and optimization. The study also examined differences in demographic and psychological variables across these profiles. Results indicated that the serial indirect effects of perceived usefulness, perceived ease of use, AI anxiety, and creative self-efficacy on usage intention via general attitude and usage attitude were significant, while their direct effects were non-significant, confirming full or partial mediation. General attitudes toward AI itself and attitudes toward using AI significantly and positively predicted usage intention in the hierarchical regression. Latent profile analysis identified three heterogeneous groups: Moderate Use Group, Low Use Group, and High Use Group. These three groups showed significant differences in grade level and key psychological variables, confirming the external validity of the profiles. The findings reveal the complexity of the psychological mechanisms underlying art design students' AI usage intention and the heterogeneous nature of their usage patterns, providing a theoretical basis and practical implications for integrating AI tools and implementing differentiated teaching strategies in art design education within the Chinese context.