The rapid expansion of artificial intelligence (AI) in international education has transformed the processes through which knowledge is produced, distributed, and governed, while simultaneously creating new challenges related to inequality, representation, and epistemic justice. This study examines algorithmic coloniality as a critical issue in AI-mediated international education, particularly within Global South contexts where technological systems are frequently shaped by dominant epistemological assumptions and unequal structures of knowledge production. Employing a qualitative conceptual research design through critical conceptual synthesis, this study integrates perspectives from artificial intelligence, international education, decolonial theory, epistemic justice, and Global South epistemologies to develop a new theoretical framework. The analysis identifies five interconnected dimensions of algorithmic coloniality: epistemic exclusion, linguistic dominance, data colonialism, digital dependency, and algorithmic governance. To address these challenges, this study develops the Decolonial Ayatutu–Ubuntu Framework, which integrates epistemic reclamation, relational ethics, algorithmic accountability, critical AI literacy, and participatory governance as foundations for more equitable AI-mediated education. The framework extends existing approaches to AI governance by shifting attention beyond algorithmic efficiency and procedural fairness toward knowledge plurality, cultural recognition, and contextual responsibility. This study contributes to global debates on responsible artificial intelligence and international education by positioning Global South epistemologies as sources of theoretical innovation rather than merely contexts affected by technological transformation, providing a conceptual foundation for rethinking AI-mediated international education beyond algorithmic efficiency toward a more inclusive, culturally responsive, and epistemically just future.