Designing a Data Governance and AI Ethics Model in the Management of Public Universities in Iran (Mixed-Methods Approach)

Abstract

Objective: This study was conducted with the aim of designing a data governance and artificial intelligence (AI) ethics model in the management of public universities in Iran. Methodology: In terms of purpose, the present research is applied–developmental, and in terms of the nature of the data, it adopts a mixed-methods design of the exploratory sequential type (qualitative–quantitative). In the qualitative phase, semi-structured interviews were conducted with 22 experts in the fields of university management, data governance, and AI ethics, and the data were analyzed using Braun and Clarke’s thematic analysis method in MAXQDA software. In the quantitative phase, a researcher-made questionnaire was developed based on the qualitative findings and distributed among 487 middle managers and senior experts of public universities in Iran. The quantitative data were analyzed using structural equation modeling with a partial least squares approach (PLS-SEM) in SmartPLS version 4. Findings: In the qualitative phase, five main themes for data governance (structural and institutional data governance, data quality and standardization, data security and privacy, data infrastructures and technologies, and a data-driven culture and human capital) and four main themes for AI ethics (transparency and explainability, fairness and algorithmic justice, human responsibility and accountability, and privacy and individual agency) were identified. In the quantitative phase, all nine research hypotheses were confirmed. The strongest effect was observed in the relationship between “data security and privacy” (β = 0.523), followed by the relationship between “data quality and AI transparency” (β = 0.478). The final model demonstrated strong goodness of fit (GOF = 0.462). Furthermore, the mediating role of transparency in the relationship between data quality and algorithmic fairness, as well as the moderating role of managerial experience in strengthening key relationships, were confirmed. Conclusion: The final model of the study consists of five dimensions governance, technical, legal, behavioral, and value-based whose structural relationships have been validated within public universities in Iran. The realization of AI ethics in public universities in Iran is not feasible without the simultaneous and systematic establishment of the five dimensions of data governance. It is recommended that the Ministry of Science develop a “National AI Ethics Framework in Higher Education,” and that universities take action to establish AI ethics committees, invest in data security and quality, and provide training in data literacy and AI ethics for managers and staff.                                                                                  

Keywords: Data governance, AI ethics, University management, Public universities in Iran, Mixed-methods approach.