Digital transformation for learner equity: The role of artificial intelligence tutors in achieving sustainable development goal 4 in Sub-Saharan Africa

Richmond Ampah Mensah (Department of Informatics, Faculty of Engineering and Computer Science, Universitas Muhammadiyah Semarang, Semarang, Central Java 50273)
Sidney Odongo (Department of Food Technology, Faculty of Agricultural Science and Technology, Universitas Muhammadiyah Semarang, Semarang, Central Java 50273)
Abdul Agbo Aziz (Department of Management, Faculty of Business and Economics, Universitas Muhammadiyah Semarang, Semarang, Central Java 50273)
Diana Hardiyant (Department of English Literature, Faculty of Education and Humanities, Universitas Muhammadiyah Semarang, Semarang, Central Java 50272)

Asian Journal Collaboration of Social Environment and Education
Vol. 4 No. 1: (July) 2026

Background: Educational imbalance continues to constrain progress toward sustainable development goal 4 (SDG 4) in Sub-Saharan Africa (SSA). Chronic teacher shortages, overcrowded classrooms, and high data costs exclude many marginalised learners and sustain high levels of learning poverty. High quality human tutoring can generate large learning gains of around one third of a standard deviation, consistent with a pooled effect size of 0.37, but remains financially and logistically difficult to scale in SSA. Methods: This article applies a comparative synthesis framework to evidence from randomised controlled trials, large scale field pilots, case studies, and policy and monitoring reports. It compares traditional human tutoring, AI supported tutoring systems, and low bandwidth mobile first interventions to assess their implications for scalability, affordability, and learner equity, and interprets the findings through economic, technological, and inclusion lenses and a decolonial perspective. Findings: Emerging AI tutors and mobile based platforms in Ghana, Sierra Leone, and Kenya indicate that low bandwidth, mobile first designs can approximate tutoring like gains while substantially reducing marginal cost and data usage, for example through WhatsApp and SMS delivery. These tools can serve as force multipliers for overstretched teachers and expand access to curriculum aligned support for learners in rural and low income communities. At the same time, persistent constraints related to infrastructure, teacher capacity, data governance, and risks of “AI in Education colonialism” limit who can benefit and how sustainably. Conclusion: AI tutors can contribute to more equitable learning in SSA when they are embedded in teacher led routines, engineered for low bandwidth environments, and governed by robust, decolonially informed frameworks that protect data and centre local curricula, languages, and communities. Their contribution depends on parallel investments in infrastructure, teacher professional development, and culturally responsive AI design. Novelty/Originality of this article: This study proposes a decolonially informed, low-bandwidth AI tutoring framework to promote equitable and sustainable education in underserved communities.