Human-centred Artificial Intelligence and Learning Analytics in emerging higher education: a critical scoping review

Abstract

Artificial Intelligence (AI) and Learning Analytics (LA) are increasingly being adopted by higher education institutions (HEIs) to support student learning, institutional decision-making, and student success. However, most published evidence originates from high-income countries, and comparatively little is known about how these technologies are implemented within the institutional, governance, and socio-cultural contexts of emerging economies. This critical scoping review synthesises the current evidence on AI and LA implementation across emerging higher education systems, with particular attention to educational outcomes, implementation contexts, equity, and institutional readiness. The review followed the methodological framework of Arksey and O’Malley [2] and was reported in accordance with the PRISMA Extension for Scoping Reviews (PRISMA-ScR). Peer-reviewed and selected grey literature published between 2019 and 2025 were systematically searched across major education, computer science, and multidisciplinary databases. Nineteen primary studies met the eligibility criteria, together with one contextual background study retained to inform interpretation of the historical development of AI and LA implementation. The evidence comprised empirical investigations, systematic and scoping reviews, policy reports, bibliometric analyses, and conceptual studies. The reviewed evidence indicates considerable regional variation in both the maturity and emphasis of AI and LA implementation. Studies from Latin America primarily describe collaborative institutional capacity-building initiatives, while evidence from Sub-Saharan Africa focuses on ethical governance, data justice, and responsible implementation. Studies from South Asia predominantly examine predictive modelling with comparatively limited evaluation of pedagogical impact, whereas evidence from East Asia and the Middle East remains exploratory and reflects early stages of institutional implementation. Across the reviewed literature, predictive early-warning systems, adaptive learning platforms, institutional analytics dashboards, and generative AI applications demonstrate potential to support educational decision-making when implemented alongside appropriate human oversight, educator involvement, and governance arrangements. However, the evidence base remains heterogeneous, with relatively few longitudinal investigations, limited causal evaluation, and substantial variation in methodological quality. The review identifies algorithmic myopia as a recurring pattern within the current literature, whereby research frequently prioritises predictive performance over educational effectiveness, learner experience, equity, and long-term institutional impact. Building on this synthesis, the paper proposes a Human-Centred and Contextually Responsive AI Framework that integrates responsible AI, equity by design, educator participation, and contextual governance to support sustainable implementation in emerging higher education systems. The review concludes with evidence-informed recommendations for higher education institutions, policymakers, technology providers, and international development partners, together with a future research agenda aligned with Sustainable Development Goal 4.