Role of AI characteristics, AI literacy, and institutional policy in shaping generative AI use and learning outcomes among higher education students

This study develops and empirically tests a comprehensive model explaining generative AI (GenAI) use among higher education students by integrating AI characteristics (perceived animacy, intelligence, and anthropomorphism) and user characteristics (self-efficacy, personal innovativeness, and AI literacy). It further examines the moderating role of institutional policy on the relationships between GenAI use, learning performance, and satisfaction. Survey data were collected from 509 university students in Oman and analyzed using a hybrid PLS-SEM and artificial neural network (ANN) approach. The results show that perceived intelligence, AI literacy, self-efficacy, and perceived anthropomorphism significantly influence GenAI use, with perceived intelligence emerging as the strongest predictor, while perceived animacy and personal innovativeness are not significant. GenAI use has a positive effect on learning performance and satisfaction. Moderation analysis reveals a significant negative interaction, indicating that stronger institutional policy increases overall learning performance and satisfaction but reduces the marginal impact of GenAI use on these outcomes. ANN results corroborate the SEM findings and further highlight perceived intelligence and AI literacy as the most influential predictors. The study advances understanding of GenAI adoption by combining sociotechnical and institutional perspectives and offers practical implications for universities and policymakers seeking to promote responsible and effective AI use in education.