Measuring Artificial Intelligence Integration in South African Private Higher Education: Applying a Refined Academic Quality Management Model

Abstract

This study applies a refined model of Artificial Intelligence (AI) integration into Academic Quality Management Systems (AQMS) within South African private higher education. It builds on prior validation through exploratory factor analysis and reliability testing to examine how institutional factors relate to stakeholders’ perceptions of AI integration. Survey data from higher education and education technology stakeholders were analysed using PLS-SEM, correlation, multiple regression, and descriptive statistics. The results show moderate explanatory power, with regression indicating stronger overall model performance than PLS-SEM. Across analyses, Research Governance and Responsible AI Use emerged as the most consistent and statistically supported factor associated with AI_QMS. Strategic Planning and AI in Teaching and Learning also contributed positively, although their effects varied by method. In contrast, AI Literacy and Competency Development and Assessment and Evaluation produced inconsistent results, including negative regression coefficients, which are interpreted cautiously due to shared variance and measurement limitations. Other factors, including Ethical Considerations and Data Privacy, Environmental Impact, and Interdisciplinary Collaboration, did not demonstrate independent explanatory power in multivariable models. The findings highlight that observable institutional activity does not necessarily translate into consistent integration within academic quality systems. The study contributes a multi-dimensional framework that distinguishes between strongly supported relationships, method-dependent effects, and areas requiring further measurement development. Practically, the model provides institutions with a structured approach to assess and strengthen AI integration within academic quality management.