Ethical Cognition, Anxiety, and Attitudes toward Artificial Intelligence in Higher Education: Validation and Predictive Modelling of the Albanian GAAIS

Artificial intelligence is increasingly embedded in higher education, shaping teaching, assessment, and institutional governance. While existing research often focuses on perceived usefulness and performance, less attention has been paid to how ethical evaluations and emotional responses jointly shape attitudes toward artificial intelligence. This study examines attitudes toward artificial intelligence in Albanian higher education through an integrated cognitive–affective–ethical framework, focusing on ethical cognition and artificial intelligence–related anxiety as key determinants of positive and negative attitudes. A cross-sectional survey was conducted among 705 students and academic staff at Albanian universities. Three validated instruments were translated and culturally adapted: the General Attitudes toward Artificial Intelligence Scale, the Attitudes toward Ethical Artificial Intelligence Scale, and the Artificial Intelligence Anxiety Scale. Confirmatory factor analyses were performed to assess factorial validity and reliability. Hierarchical multiple regression models were then estimated to examine the predictive roles of ethical attitudes and artificial intelligence–related anxiety on positive and negative attitudes, controlling for sociodemographic characteristics. Confirmatory factor analyses supported the original factorial structures of all three instruments, with excellent model fit and strong internal consistency. Ethical cognition and emotional responses jointly explained substantial variance in attitudes toward artificial intelligence. Among ethical dimensions, non-maleficence was the only factor consistently associated with more positive attitudes, whereas privacy concerns significantly increased negative evaluations. Both cognitive and affective forms of artificial intelligence anxiety reduced positive attitudes and amplified negative ones, with affective anxiety showing the strongest effects. Ethical and emotional factors together explained 47% of the variance in positive attitudes and 39% in negative attitudes. Attitudes toward artificial intelligence in higher education are shaped by an interplay of ethical evaluation and emotional response rather than by instrumental considerations alone. Ethical reassurance related to harm avoidance and effective management of anxiety appear central to fostering constructive engagement with artificial intelligence. These findings highlight the importance of integrating ethical reflection and emotional preparedness into artificial intelligence literacy initiatives, particularly in transitional higher-education contexts.