Convergence and divergence in the discourse of AI governance: a comparative analysis of global and African contexts

This study examines convergence and divergence between Africa-focused and global-level AI governance narratives. Using sentence-level embeddings and K-means clustering, the analysis identifies five thematic clusters. Whereas the selected global-level documents place greater emphasis on safety, risk management and multilateral coordination, Africa-focused documents foreground sovereignty, regional governance and capabilities required to develop and deploy AI. These differences are consistent with different developmental priorities and unequal technological and institutional positions within the global AI ecosystem. The prominence of sovereignty and capacity-building, together with African engagement across multiple governance venues, is more consistent with context-sensitive strategic participation than with passive rule-taking. The findings highlight the limitations of universal governance frameworks and underscore the need for parallel development of technological capabilities, governance institutions and normative safeguards.