This research investigates the paradoxical dual-use nature of Artificial Intelligence (AI) in Africa’s anti-corruption efforts, where AI-enabled whistleblowing systems (AI-EWS) can serve either as empowering “algorithmic guardians” for citizens or as repressive “digital snitches” for the state. Through a comparative case study of South Africa, Kenya, and Ethiopia, and the development of a novel AI-Sousveillance-Paternalism (AI-SP) Framework—which synthesizes theories of sousveillance and technological paternalism—the study analyzes how a system’s orientation is configured by its architectural design, data governance, and institutional routing. Findings from document analysis and stakeholder interviews demonstrate that polycentric routing and transparent architecture, within conducive political environments, foster civic trust and bottom-up accountability (sousveillance). In contrast, opaque, centralized systems reinforce state control and paternalism. The study concludes that AI’s transformative potential for accountability operates as a catalytic amplifier of existing institutional dynamics rather than as an independent causal force. It offers critical design and policy principles centered on distributed trust, data sovereignty, and algorithmic transparency to ensure these technologies function as tools of public empowerment rather than instruments of repression.
Walle et al. (Mon,) studied this question.