Fighting Learning Poverty Using AI.

Bridging Learning Poverty with Artificial Intelligence: Equity Imperatives for African Education.

Introduction

Learning poverty is defined as the inability of a child to read and understand a simple text by age ten, remains a persistent challenge across many African countries (World Bank, 2019). Chronic shortages of instructional resources, overcrowded classrooms, teacher workload pressures, and limited early literacy interventions continue to contribute to widespread foundational skill deficits (Spaull & Pretorius, 2019; UNESCO, 2022). At the same time, the rapid expansion of artificial intelligence (AI) is reshaping global education systems by offering adaptive learning solutions, personalized feedback, and new pathways for inclusive instruction (Luckin, 2018; OECD, 2021).

The potential of AI to reduce learning poverty in Africa is significant, particularly in the areas of early literacy and individualized learning support. However, AI adoption is shaped by deep structural inequalities including digital infrastructure disparities, language exclusion, socio-economic divides, and limited teacher professional development—that risk widening existing gaps if not deliberately addressed (Birhane, 2020; UNESCO, 2021). This article extends earlier work examining learning poverty in Ethiopia (Yirgalem, 2022) by situating AI adoption within a broader, continent-wide equity discourse. The central argument advanced here is that AI must be designed and implemented to advance, rather than compromise, educational justice.

Theoretical Framework

This analysis is guided by two complementary theoretical lenses: the Capability Approach and Critical Technological Equity frameworks. The Capability Approach foregrounds learners’ real freedoms and opportunities to achieve valued educational outcomes, emphasizing not only access to education but also the quality and relevance of learning experiences (Sen, 1999; Nussbaum, 2011). In African contexts where structural barriers limit access to high-quality instruction, AI has the potential to expand learners’ capabilities by personalizing instruction, supporting multilingual literacy development, and enabling differentiated feedback (Vaughan et al., 2023).

Critical Technological Equity frameworks emphasize that technologies are never neutral; rather, they are shaped by existing political, social, and economic inequalities (Benjamin, 2019; Eubanks, 2018). In educational settings, these inequalities manifest through algorithmic bias, exclusion of local languages, and uneven access to digital infrastructure. These inequalities affect disproportionately marginalized learners in Africa (Mutonyi & Kizito, 2022). Together, these frameworks underscore the necessity of designing and implementing AI systems that are context-responsive, culturally relevant, and equitably distributed if they are to meaningfully address learning poverty.

AI Opportunities and Equity Challenges

Artificial intelligence offers substantial opportunities to enhance learning outcomes across African education systems. Adaptive learning platforms can tailor instruction in real time, enabling large classrooms—often characterized by student–teacher ratios exceeding 50:1 to provide more individualized learning pathways (Trucano, 2021; UNESCO, 2023). Research indicates that AI-enhanced early literacy tools can support phonemic awareness, vocabulary development, and reading fluency, particularly when delivered through mobile technologies accessible to families and communities (Heugh, 2021; World Bank, 2022). AI-supported diagnostic analytics also enable educators to identify learning gaps early, strengthen formative assessment practices, and prevent long-term academic deficits (Holmes et al., 2019; OECD, 2021). Additionally, assistive technologies powered by AI such as text-to-speech and speech-to-text tools provide critical support for learners with disabilities and align with Universal Design for Learning principles (Al-Azawei et al., 2017; UNICEF, 2021).

Despite these opportunities, significant equity challenges remain a major concern. The digital divide characterized by limited broadband connectivity, device shortages, and unreliable electricity continues to restrict meaningful access to AI-enabled learning across rural and underserved communities (GSMA, 2022; World Bank, 2023). Furthermore, AI systems developed primarily using Western datasets often fail to reflect African languages and cultural contexts, resulting in inaccurate, biased, or culturally irrelevant outputs (Birhane, 2021; UNESCO, 2021). Ethical concerns, including data privacy risks, child surveillance, and weak regulatory frameworks, further heighten vulnerabilities in contexts lacking robust data protection policies (African Union, 2022; Hintz et al., 2020). Teacher readiness also poses a critical challenge; without sustained professional development, educators may lack the confidence and pedagogical grounding necessary to integrate AI tools effectively into classroom practice (Miao & Holmes, 2021; Oketch & Rolleston, 2020). These intersecting challenges highlight the need for deliberate, equity-oriented AI implementation strategies.

Discussions

Effectively addressing learning poverty through AI requires coordinated, system-wide reform. First, AI must be positioned as a tool that supports, not replaces teachers. Evidence consistently demonstrates that teacher-led, AI-supported instructional models produce the strongest learning outcomes, particularly in literacy and numeracy (Luckin et al., 2016; Schleicher, 2020). Professional learning initiatives should therefore focus on building teachers’ digital competence, ethical awareness, and critical judgment in selecting and applying AI tools.

Second, linguistic inclusion is essential for equitable learning outcomes. Research shows that early literacy development is strongest when children learn to read in their mother tongues; however, the majority of AI educational tools do not adequately support African languages (Heugh, 2021; UNESCO, 2023). Developing AI models trained on African linguistic datasets is therefore central to reducing early literacy gaps. Third, sustained investment in infrastructure remains non-negotiable. AI cannot meaningfully reduce learning poverty in contexts where schools lack reliable electricity, stable connectivity, or affordable devices (AUDA-NEPAD, 2021; World Bank, 2023). Governments must prioritize infrastructure development in rural and marginalized communities to ensure equitable access.

Fourth, ethical governance frameworks are essential to protect children’s rights. Clear national guidelines addressing data protection, transparency, and algorithmic fairness must accompany AI adoption in education systems (African Union, 2022; UNICEF, 2021). Finally, sustainable AI integration will require strong partnerships among governments, universities, teacher education institutions, and African technology developers to build scalable, culturally grounded innovations (Tadeu & Ferreira, 2021).

Conclusion and Recommendations

Artificial intelligence holds considerable potential to address learning poverty across Africa, but only when implemented through strategies centered on equity, linguistic relevance, and ethical governance. To realize this potential, African governments should invest in digital infrastructure, expand teacher professional development in AI literacy, and establish robust national frameworks for ethical and child-sensitive AI use (African Union, 2022; World Bank, 2022). AI developers must design tools grounded in African languages, cultural contexts, and pedagogical priorities to ensure relevance and fairness (Birhane, 2021; UNESCO, 2021). Research institutions and universities should lead rigorous evaluations of AI tools to assess their effectiveness in reducing learning poverty and advancing educational equity. By situating AI adoption within broader systemic reforms and prioritizing equity at every stage, African education systems can leverage AI not only to close learning gaps but also to build inclusive, empowering pathways for learners’ futures.