Data Quality and the Trust–Interpretability Paradox: Toward a Mid-Range Framework for AI Adoption in Educational Information Systems
, Bassam Al-Shargabi, Osama Nashaat AttiaThis study examines limitations in existing AI adoption models in education, which often treat data quality, interpretability, trust, and organizational factors as independent elements. This separation may lead to incomplete explanations of AI implementation outcomes in educational information systems.
This study introduces and empirically examines the Data–Knowledge Alignment Theory for Educational Information Systems (DKAT-EIS). The framework draws on insights from Information Systems, Knowledge Management, and Explainable Artificial Intelligence to explore how data quality, knowledge interpretability, and trust relate to AI adoption in universities. DKAT-EIS is presented as a context-specific analytical framework that offers initial insights into these relationships in higher education.
A quantitative survey was conducted with 1,150 respondents (students, faculty, and staff) at Jazan University in Saudi Arabia. Cross-sectional data were analyzed using Structural Equation Modeling (SEM) and Confirmatory Factor Analysis (CFA) to test the proposed relationships.
The study proposes and examines the DKAT-EIS framework to better understand how data and knowledge processes influence AI adoption in higher education. The findings highlight the role of data quality in supporting knowledge interpretability and indicate that the relationships between interpretability, trust, and adoption may be more complex than suggested in traditional technology adoption models. Given the single-institution and cross-sectional design, the findings should be interpreted as context-specific evidence that encourages further validation in other settings.
Data quality significantly predicts knowledge interpretability. However, interpretability does not significantly influence ethical trust, and trust does not directly predict AI adoption success. This pattern suggests a “trust–interpretability paradox.” In addition, organizational readiness does not significantly moderate the relationships in the model.
Educational institutions should prioritize data governance and data quality when implementing AI systems. Improving system transparency alone may not build trust; organizations should also address ethical, organizational, and cultural factors that strengthen user confidence.
Future studies should examine additional factors, such as fairness perceptions, digital culture, and institutional context, that may explain the gap between interpretability and trust. Further testing of the DKAT-EIS framework across different universities and regions is recommended.
The findings suggest that responsible AI adoption in higher education requires a balanced socio-technical approach that combines strong data foundations with attention to ethical and organizational factors that foster trust.
Future research should include longitudinal studies, cross-institutional validation, and qualitative investigations to better understand how contextual factors influence AI adoption, trust, and interpretability.
Back