Artificial Intelligence, has become a transformative force in education, enabling personalized learning, adaptive assessment, and data-driven pedagogy. Yet, its integration poses significant risks, notably algorithmic—particularly cultural—bias, which can exacerbate systemic inequities. This study analyzes the origins, manifestations, and educational impacts of such bias, drawing on literature from computer science, education, ethics, and law. It highlights how unrepresentative datasets, biased model architectures, and sociocultural blind spots lead to discrimination in assessment, learning recommendations, and resource allocation. Using a critical synthesis of empirical research and policy analysis, supported by international case studies, the study finds that cultural underrepresentation, opaque decision-making, and weak governance frameworks undermine fairness and equity in AI-driven education. Such biases distort performance evaluations, reinforce stereotypes—such as gendered career guidance in STEM— and widen disparities. The paper recommends diverse datasets, transparent and explainable AI (XAI), institutionalized fairness audits, and the integration of ethical AI principles in education. It underscores the role of educators, policymakers, and international bodies in establishing accountability, inclusivity, and cultural adaptability. Achieving equitable AI-supported education demands sustained interdisciplinary collaboration, combining innovation with robust ethical governance.
HOCA et al. (2025) studied this question.