From Data Colonialism to Inclusive Artificial Intelligence: Governing Digital Transformation in Sub-Saharan African Food Systems

Abstract

Artificial intelligence (AI) is increasingly promoted as a transformative technology for improving agricultural productivity, climate resilience, and food security in Sub Saharan Africa (SSA). However, growing evidence suggests that the governance of AI often reproduces existing inequalities by concentrating control over agricultural data, decision making, and economic value among external actors. This paper examines AI governance in SSA food systems through the interconnected lenses of data power, equity, and digital transformation. Agricultural AI systems rely on data generated through farming activities, management practices, production records, remote sensing, Internet of Things devices, drones, satellites, and other digital technologies. Although these datasets originate from farmers and their production environments, rights over data access, ownership, processing, reuse, and commercialization are frequently exercised by platform providers, technology firms, and development organizations, raising fundamental questions regarding data sovereignty, participation, and benefit sharing. The study draws on a semi systematic literature review with narrative synthesis of 92 interdisciplinary studies published between 2015 and 2025. Guided by a political economy perspective integrating theories of data colonialism and platform capitalism, the analysis develops a governance framework based on three analytical dimensions: data control, participation, and value distribution. These dimensions are used to conceptualize a governance continuum ranging from extractive to inclusive AI systems, providing a comparative framework for analysing governance trajectories across SSA. Evidence from Kenya, Rwanda, Nigeria, Ethiopia, Tanzania, Uganda, and South Africa shows that contemporary AI deployments are predominantly characterized by centralized data control, limited farmer participation, and unequal value capture, although important institutional variation exists across countries. The review also identifies emerging alternatives, including participatory design, farmer data cooperatives, and locally embedded innovation ecosystems, which demonstrate pathways toward more inclusive AI governance but remain fragmented and insufficiently institutionalized. The findings highlight persistent policy gaps in agricultural data sovereignty, sector specific regulation, and institutional capacity, and argue that the transformative potential of AI will depend less on technological innovation than on governance arrangements that redistribute power, strengthen local agency, and ensure equitable sharing of the benefits generated through AI enabled food systems. This study contributes to a transferable governance continuum that advances comparative analysis of AI governance and offers a practical framework for designing more inclusive digital agriculture policies in Sub Saharan Africa and comparable developing regions.