Artificial Intelligence Sovereignty in Africa: Legal, Ethical and Strategic Challenges and Policy Pathways

Abstract

Artificial intelligence (AI) has emerged as a decisive terrain of twenty-first-century geopolitical competition, and Africa's engagement with this terrain raises urgent questions about sovereignty, self-determination, and equitable participation in global technological governance. This paper examines the legal, ethical, and strategic challenges to AI sovereignty in Africa, defined as the continent's capacity to govern, own, and direct the development and deployment of AI technologies within its territories according to its own values and developmental priorities. The analysis is based on a qualitative, doctrinal and interpretive desk-based review of six primary continental and national instruments which principally include the African Union's Continental Artificial Intelligence Strategy (2024), Nigeria's National Artificial Intelligence Strategy (2025), and the Nigeria Data Protection Act (2023)  read alongside twelve peer-reviewed and policy publications by African and diaspora scholars, published between 2020 and 2025 and identified through targeted searches of Google Scholar, SSRN and the institutional repositories of the African Union, Research ICT Africa, and the Strathmore University Centre for Intellectual Property and Information Technology Law (search terms and selection criteria are set out in full in Section 4). The paper's principal contribution is an integrated analytical model that treats African AI sovereignty as constituted by the interaction of four dimensions which are legal authority, ethical authority, infrastructural capacity, and epistemic standing  rather than as a single attribute a state either possesses or lacks. Applying this model, the paper argues that Africa's AI sovereignty is constrained by fragmented and unevenly enforced legal regimes, infrastructural and compute dependency on external powers, a persistent brain drain of AI talent, and epistemic marginalisation within global AI ethics discourse, while showing how African institutions and scholars are actively reconstructing sovereignty through relational ethical frameworks such as Ubuntu, harmonised data-governance proposals, and assertive continental strategy documents. As a desk-based doctrinal and interpretive review rather than an empirical study, the paper's principal limitation is that its conclusions are interpretive readings of a purposively selected corpus rather than statistically representative findings across all fifty-five African states. The paper closes with prioritised, actor-specific recommendations for the African Union, regional economic communities, national governments with particular reference to Nigeria covering regulators, universities, and private-sector actors.