<![CDATA[Artificial intelligence (AI) is rapidly expanding across African societies, yet most AI ethics frameworks derive from Western individualistic philosophies. Indigenous knowledge systems (IKS), including Ubuntu, Ifa, and other ancestral epistemologies offer alternative, relational foundations for ethical and effective AI. This systematic review synthesizes peer reviewed and grey literature on AI-IKS integration in African societies, identifying applications, ethical frameworks, governance models, and documented outcomes. Following PRISMA 2020 guidelines, we searched Scopus, Web of Science, Google Scholar, and African Journals Online (2010-2025). Thirty four studies met inclusion criteria. Thematic synthesis was conducted after quality appraisal using the Mixed Methods Appraisal Tool and a theory focused framework. Four themes emerged: (1) Ubuntu is widely proposed as a relational AI ethics alternative, but operationalization is minimal (only two prototypes). (2) Practical applications exist in agriculture (ITIKI drought prediction, 78% accuracy), health (AI medicinal plant mining), language, and architecture, but only three systems are field validated. (3) Risks of epistemic extraction, data sovereignty violations, and stereotype perpetuation are well theorized yet rarely mitigated. (4) Governance is fragmented; no African country has a dedicated IKS AI policy. The field is conceptually rich but empirically thin. A schism exists between Ubuntu ethics discourse and deployed systems. Future research must prioritise community led co design, enforceable data sovereignty, longitudinal impact assessment, and publication of negative results.]]>
Copyrights © 2026
