Higher education transformation through artificial intelligence: an empirical study in Saudi Arabia

This study examines associations between AI awareness, readiness, ethical perceptions, attitude, and actual AI usage.Survey data from 424 students at King Fahd University of Petroleum and Minerals (KFUPM) were analyzed using Partial Least Squares Structural Equation Modeling (PLS-SEM) in SmartPLS 4. Two structural models were compared: one specifying direct effects and a second in which attitude served as a partial mediator. In Model 1 (direct effects), only readiness and attitude were significantly associated with AI usage. In Model 2 (mediation), readiness retained both direct and indirect associations with usage via attitude; ethics displayed a competitive mediation structure, simultaneously building positive attitudes that supported usage while directly restraining use when ethical concerns remained unresolved ; and awareness remained non-significant across all paths. These findings advance adoption theory by repositioning awareness as a distal antecedent in high-penetration contexts, elevating readiness to a central dual-pathway antecedent, and providing the first structural evidence of competitive mediation for ethical perceptions in a higher education AI adoption model. For institutions, the findings indicate that building structured student capability and embedding transparent ethical governance are the two essential conditions for AI adoption that are not only widespread but also responsible and sustainable.