Higher Education Students’ Trust Formation and Behavioral Intention Toward GenAI in Interdisciplinary Learning: A Behavioral Reasoning Theory Perspective

Generative artificial intelligence (GenAI), with its capacity to integrate multidisciplinary knowledge and address complex problems, has been widely applied in interdisciplinary learning among higher education students. However, its reliability remains controversial due to issues such as hallucination and privacy concerns, leading students to oscillate between trust and skepticism in actual use. To examine the mechanisms underlying GenAI trust formation and its influence on behavioral intention, this study adopts Behavioral Reasoning Theory and a hybrid analytical approach combining partial least squares structural equation modeling (PLS-SEM) and fuzzy-set qualitative comparative analysis (fsQCA). The PLS-SEM results indicate that perceived usefulness, information accuracy, and information clarity positively influence trust, whereas hallucination risk and interdisciplinary context insensitivity hinder trust formation. Autonomous learning strengthens reasons for using GenAI while partially suppressing reasons against its use, except for privacy concern. Furthermore, perceived usefulness, information clarity, and trust have significant positive effects on behavioral intention. The fsQCA findings reveal that information clarity and low hallucination risk are core conditions for high trust, and perceived usefulness and trust are central to high behavioral intention. Notably, one configuration shows that trust in GenAI coexists with concerns about hallucination risk, suggesting that users’ willingness to use GenAI reflects a compromise rather than the absence of risk concerns. Overall, this study provides a nuanced understanding of the psychological mechanisms underlying technology trust and adoption and offers practical implications for the application of GenAI in interdisciplinary learning.