AI is no longer a distant or abstract policy issue. It is becoming part of the infrastructure through which states govern, firms compete, and citizens access services. Across the world, states are developing national AI strategies, international bodies are debating governance frameworks, and private actors continue to expand the use of AI in everyday decision-making. In that context, much of the discussion has focused on fairness, transparency, accountability, and explainability. Those principles remain important. Yet for Africa, they are not enough on their own.
The deeper question is not only whether AI systems are fair in a technical sense. The question is whether they will reinforce existing inequalities or help correct them. That is why justice should shape Africa’s AI future. A fairness-centred approach often asks whether a system treats cases similarly or whether it reduces measurable bias. A justice-centred approach asks broader questions: who develops the technology? Who owns the infrastructure? Who controls the data? Who captures the value? And who is excluded from participation? Those are not secondary issues. For Africans, they are the core of the AI governance challenge.
On the continent, stakes are particularly high. AI is being adopted within an already unequal global digital order. In its 2024 report “Governing AI for Humanity,” the UN Secretary-General’s High-Level Advisory Body on Artificial Intelligence (HLAB-AI) found that 118 countries remain excluded from the major initiatives shaping AI governance globally, despite being directly affected by its outcomes. Many of those countries are in Africa, where states still depend on imported technologies, external platforms, and infrastructure they do not fully control. Africa is not entering the AI era from a position of equal bargaining power, yet it is often expected to adapt to systems designed elsewhere, under assumptions that may not reflect its legal, economic, linguistic, or social realities.
That imbalance matters. If African institutions are only users of AI rather than active shapers of AI, the continent risks becoming a passive consumer of technologies that do not reflect its own priorities. Governance frameworks developed elsewhere may address certain technical risks, but they may not adequately address structural questions of development, access, and power. Justice, therefore, becomes more than a moral ideal. It becomes a necessary framework for evaluating whether AI is serving African societies or simply reproducing dependency in a new form.
This is where Research ICT Africa’s Africa Just AI programme provides an important contribution. The programme’s work across data governance, intellectual property, economic policy, cybersecurity, democracy, gender, and green energy shows that the risks associated with AI are not limited to algorithmic bias. They also arise from the conditions under which AI is produced and deployed. Who has access to computational infrastructure? Whose knowledge is represented in the training data? Who has the ability to influence standards and regulation? These are questions about power as much as they are about technology.
A justice-based approach also helps clarify why the global conversation on AI can sometimes feel incomplete. Much of the policy debate assumes that once a system is built, the main task is to make it transparent, accountable, and auditable. The ITU’s 2024 AI Governance Day report, titled “From Principles to Implementation,” identifies exactly this gap as the central challenge for the next phase of global AI governance. In Africa, the problem runs deeper still. The more urgent question may be whether a system should have been designed in its current form at all, and whether the surrounding social conditions make its deployment equitable. Governance should begin well before the point of technical compliance.
This point is especially relevant in relation to digital infrastructure. AI depends on electricity, data connectivity, cloud systems, institutional capacity, and reliable public services. If those foundations are weak, AI will not produce broad social benefit, no matter how sophisticated the algorithms may be. In African cities, where governments are increasingly exploring AI in land administration, transport systems, service delivery, and urban planning, this issue becomes even more threatening. AI can help improve efficiency, but it can also deepen spatial and social inequalities if deployed in settings where access is uneven and institutional safeguards are limited.
The urban context also makes the justice question impossible to ignore. Cities are not neutral spaces. They reflect histories of exclusion, uneven development, and unequal access to land, housing, transport, and services. If AI is introduced into those environments without a justice-oriented governance framework, it may simply automate existing patterns of exclusion. A system that appears efficient may still reproduce unfair outcomes if it is built on unequal data, biased assumptions, or weak institutional oversight.
Participation is another important concern. UNESCO’s Recommendation on the Ethics of Artificial Intelligence, adopted in 2021 by all 193 UNESCO Member States, recognises inclusive multi-stakeholder governance as a foundational principle, not an optional feature. Yet for Africa, meaningful participation in the global structures that shape AI norms and standards remains limited. African researchers, regulators, and innovators are still underrepresented in the spaces where those decisions are made. That has consequences. Many of the frameworks eventually applied in African contexts are shaped by priorities developed elsewhere. A justice-based approach challenges that imbalance by insisting that African perspectives are not added at the end of the conversation, but built into it from the beginning.
This is why Africa should not simply import governance models designed in other settings. It should adapt them, question them, and where necessary, rework them entirely. The African Union’s (AU) Continental AI Strategy, endorsed by the AU Executive Council in Accra in July 2024, already charts this direction. It frames AI governance around inclusive digital transformation, human capital development, and continental agency rather than external compliance, setting out a roadmap built on Africa’s own developmental priorities. That is the right instinct. Justice gives us a way to extend it further, moving the discussion beyond narrow risk management and toward a broader evaluation of public value.
Does AI expand opportunity, strengthen institutions, support inclusive development, or allow African societies to shape their own digital futures? If not, these are the standards that should guide AI governance on the continent. The goal should not be to reject AI, nor to treat it as inherently beneficial. The goal should be to ensure that it is governed in ways that are socially legitimate, institutionally grounded, and developmentally meaningful. In Africa, that means putting justice ahead of a thin notion of fairness.
If AI is to contribute to the continent’s future, it must do more than avoid harm. It must help expand access, improve public institutions, and support more equitable forms of development. That is a higher standard, but it is the right one – the one that Africa should insist on.
