Artificial Intelligence and Digital Agriculture for Smallholder Farming: An Africa-Centred Scoping Review of Global South Evidence

Artificial intelligence (AI) is increasingly applied in agriculture, but evidence concerning its relevance to smallholder farming remains heterogeneous and geographically uneven. This scoping review mapped evidence on AI and related digital agricultural technologies, their applications, adoption conditions, and equity and governance implications for smallholder farming in Africa-centred Global South contexts. This review was reported using PRISMA-ScR guidance. Across Google Scholar, Web of Science and Scopus, 150 potentially relevant records were retained for the documented screening process, and 42 sources were included. Sources were classified as AI-specific evidence, AI-enabled systems or contextual digital-agriculture evidence and synthesised thematically. The literature addressed crop monitoring, yield prediction, irrigation, climate-risk management and digital advisory services. However, it was dominated by African evidence and included empirical studies, technical demonstrations, reviews, policy analyses and conceptual papers. The mapped sources more often described technological feasibility and adoption conditions than sustained effects on productivity, income, resilience or farmer welfare. Recurring constraints included cost, connectivity, digital literacy, language, trust and institutional capacity. Equity and governance concerns included data rights, gender inclusion, algorithmic bias and unequal benefit distribution, although many were conceptually rather than empirically documented. The findings support context-sensitive, farmer-centred implementation while highlighting the need for longitudinal and geographically broader research.