Mirʾāh

Abstract

Adaptive learning platforms now make continuous instructional decisions for millions of learners, yet those decisions remain largely invisible to the students, teachers, and families they affect. This conceptual article examines the ethical implications of that invisibility and proposes a practical response. Drawing on recent empirical research and on the systematic review literature, the analysis organizes the risks of adaptive learning into three connected dimensions, technology, education, and society, and maps how each risk falls unevenly across six stakeholder groups. The article then introduces Mirʾāh, a synthesized ethical framework of seven principles for transparent adaptive learning organized in a three tier governance architecture, a Foundation tier in which transparency and contestability create trustworthy interaction, a Protection tier in which fairness, privacy, and accountability safeguard learners and institutions, and an Empowerment tier in which educational integrity and AI literacy ensure that AI supports authentic learning and informed participation. The framework aligns with UNESCO, OECD, and European Union standards and extends them to confront the datafication of education. Finally, the article situates the framework within the policy environment of the United Arab Emirates, where a national AI strategy, a national AI charter, and a new school AI curriculum create favorable conditions for implementation and offers five governance recommendations for institutions. The framework contributes a bridge between high level international principles and the daily classroom realities of algorithmic personalization.