Across Africa, artificial intelligence is moving from experimentation into policy. As of April 2026, the Organisation for Economic Co-operation and Development (OECD) identified 17 African countries with national AI strategies, alongside others developing policy frameworks. Earlier in 2024, the African Union adopted its Continental Artificial Intelligence Strategy, along with the developments from Regional Economic Communities – the regional groups of African states – which are as follows:
- The East African Community (EAC) is developing an AI Strategy for 2026–2031, and
- Common Market for Eastern and Southern Africa (COMESA) has begun consultations for a Regional AI Strategy in August 2026.
As Africa’s AI governance evolves, an important question remains: is gender surviving this AI policy boom? The question is not simply whether the words “women” or “gender” appear in AI strategies. It is whether decades of African gender policy, institutions, and feminist advocacy are actually shaping decisions about data, infrastructure, procurement, skills, investment, and AI regulation. There is also a question beneath much of this policy conversation: what does gender diversity actually mean? Across many AI frameworks, gender inclusion is still largely understood through a binary lens, with “gender” often functioning as shorthand for women and girls. Addressing the historical exclusion of women is critical, but AI governance must also ask how AI systems classify, represent or exclude people whose identities do not fit neatly within binary categories.
Continental AI strategy
The 2024 Continental AI Strategy includes a dedicated focus on gender equality, inclusion, and diversity; calls for women and girls to benefit from AI development; AI entrepreneurship; and recommends impact assessments capable of identifying harms. This is significant because it moves beyond asking how many women work in technology. But the policy disconnect remains visible. If gender mainstreaming means integrating gender across all areas of policymaking, gender cannot sit primarily within an inclusion section. It must also shape decisions about infrastructure, data governance, AI investment, labour transitions, public procurement, transparency, rights-based impact assessment and due diligence, and accountability.
This creates an important distinction between including women in AI and governing AI through a gender lens. The former may produce scholarships, training programmes, and entrepreneurship opportunities. The latter asks whose data trains systems, whose labour is displaced, who controls infrastructure, who benefits from investment, and who has access to remedy when AI discriminates. It also requires interrogating what policies mean when they use the word “gender”. Much of the current language remains centred on women and girls rather than engaging with gender as a more complex system of identities, norms, and power relations. Few AI strategies appear to ask how automated systems may treat people whose identities fall outside binary gender classifications.
Gender cannot sit primarily within an inclusion section. It must also shape decisions about infrastructure, data governance, AI investment, labour transitions, public procurement, transparency, rights-based impact assessment and due diligence, and accountability.