CIPESA analysis warns African AI laws risk replicating Global North models
CIPESA published an analysis arguing that AI governance frameworks in Africa are often drafted without meaningful consultation and risk copying ill-fitting global-North models.
Summary
An analysis by Paul Kimumwe for CIPESA argues that African countries developing AI laws and strategies risk copying models from the global North. It first recounts the international and continental setting, noting a March 2024 UN resolution on safe, secure and trustworthy AI, the August 2025 UN resolution 79/325 creating a scientific panel and a global dialogue on AI governance, and the AU's Continental AI Strategy of July 2024.
The piece lists national efforts, saying Rwanda was first, with a national AI policy in 2019, and that Ghana (October 2022), Egypt (January 2025) and Kenya (May 2025) came later. It names several more countries with policies and others still drafting theirs. Kimumwe then argues that adopting such instruments is only a first step, because drafting in many developing contexts often lacks real multistakeholder consultation and tends to replicate global North models that can clash with local contexts and cultural norms. As an example he cites reservations several African states voiced about the Maputo Protocol.
The analysis says these models are usually state-centric and set a binary between duty-bearers and rights-holders without explaining what each must do. It points to slow ratification, noting the Malabo Convention took almost nine years to enter into force and that 40 countries have data protection laws while 16 have ratified the continental convention.
Kimumwe concludes that countries should pass AI-specific laws, strengthen oversight bodies, secure meaningful stakeholder participation and adopt a human rights-centred approach, calling the current stage an opportunity for Africans to shape AI on their own terms.