How and when does artificial intelligence use enhance student creativity? The mediating role of knowledge-based dynamic capabilities.

As artificial intelligence (AI) becomes more deeply integrated into higher education, understanding its effects on student creativity has become increasingly important. However, the mechanisms through which AI use is associated with student creativity, as well as the boundary conditions of this relationship, remain insufficiently understood. Drawing on the knowledge-based dynamic capabilities (KBDC) framework, this study develops and tests a model that links AI use to student creativity while incorporating the moderating roles of AI self-efficacy and AI trust. Using survey data from 764 university students in China, we employ hierarchical regression analysis to examine the proposed model. The results show that AI use is positively associated with student creativity and that this relationship is mediated by students' knowledge acquisition, absorptive, and combination capabilities. Furthermore, AI self-efficacy strengthens the positive effects of AI use on all three capability dimensions, whereas AI trust weakens the effects on knowledge acquisition and absorptive capabilities. In doing so, this study advances research on AI use and student creativity by unpacking one important capability-based mediating pathway and the boundary conditions under which this relationship is stronger or weaker. It also contributes to the KBDC literature by extending the application of KBDC theory from the organizational to the individual level and demonstrating how these capabilities operate in AI-enabled learning contexts. The findings offer practical implications for educators, universities, students, and policymakers seeking to foster creative and reflective AI-supported learning.