Governing artificial intelligence for educational justice: developing a framework for inclusive and ethical digital learning

Purpose This study aims to examine how artificial intelligence (AI) can be governed to advance educational justice, inclusion and ethical digital learning. It addresses incomplete integration across AI-in-education, AI-governance and educational-justice scholarship by developing a justice-oriented governance framework. Design/methodology/approach The study uses conceptual theory-building through an integrative synthesis of 186 sources identified through database searching and citation tracking for 2015–2025, with 2026 literature used only as contextual updates. The synthesis compares concepts, governance functions and theoretical relationships across domains. Findings The synthesis conceptualizes the AI–Justice Paradox as the possibility that expanded AI-enabled educational opportunity may coexist with or intensify injustice. It develops the educational justice governance framework (EJGF), comprising equitable access, capability development, algorithmic fairness, participatory governance and accountability and rights protection. These dimensions are theorized as conditionally interdependent and partially non-substitutable. Originality/value The contribution is integrative and relational: the EJGF complements established cross-sector AI-governance approaches by specifying educational justice as the substantive criterion linking governance safeguards within educational institutions. The framework is conceptually derived and requires empirical validation across contexts, institutions and AI applications. It also generates testable propositions concerning configurational relationships and boundary conditions.