Adoption of Sustainable and Human-Centered Explainable AI in Higher Education. An Extended UTAUT Perspective

Abstract

As Artificial Intelligence permeates higher education, concerns regarding its "black-box" nature, ethical implications and environmental sustainability have intensified. This study develops and empirically validates an Extended Unified Theory of Acceptance and Use of Technology (UTAUT) framework to investigate the factors influencing AI adoption among students in Zimbabwean universities. The model integrates five contemporary AI constructs which are perceived explainability, trust in AI, perceived human-centeredness, perceived sustainability and perceived fairness into the traditional UTAUT1 framework. Data were collected through a structured Likert-scale questionnaire from a sample of 352 students across major Zimbabwean HEIs. Regression analysis was employed to examine the relationships among constructs. The extended model explains 68.4% of the variance in behavioural intention (R² = 0.684). Results indicate that while performance expectancy and trust in AI are the strongest predictors, perceived explainability and human-centeredness significantly enhance adoption intentions. Surprisingly, perceived sustainability emerged as a nascent but significant driver, reflecting a growing awareness of "Green AI." The findings provide critical theoretical contributions by bridging the gap between instrumental adoption drivers and human-centric design. Practically, the study offers a roadmap for university administrators and AI developers in developing countries to foster transparent, fair and sustainable AI ecosystems.