Resilience and Preparedness in AI Governance: The African Perspective

For decades, historians have been wrestling with the concept of technology determinism – that technology is the prime factor in shaping our lifestyles, values, institutions, and other elements of our society. This doctrine about the omnipresence of technology is more important in the ongoing moment of the rapid evolution of AI technologies and the ongoing conversations on prioritising safety especially in the context of agentic AI capabilities and its impact on the future of humanity.

Some view innovation as opening a neutral door that we freely choose to enter (as Lynn White famously suggested), but technology often carries its own biases and directions as the historian, Kranzberg, rightly points out, who decides which doors to open and when we enter that door, are our future actions not guided by the contours of the corridor and can we really turn around once we enter the door?

These are important questions to answer particularly in the context of Africa’s place in the global AI economy and how we build economic resilience. In late 2024, I had the opportunity to attend the Rhodes Forum on Technology and Society at which Patrick Pichette, the former CFO of Google who was also at the helm when Google rolled out defining innovations of our time such as Google Maps, delivered a keynote address. In it, he described technology as inherently neutral with the phrase – “technology: it just is!” Pichette was championing technology as an inevitable part of modern life and a force for good. But this claim is in direct contradiction to Kranzberg’s first law of technology which states that “technology is neither good nor bad; nor is it neutral,” because its social effects depend on how and by whom it is used. Applying this ‘first law’ to Africa’s engagement with AI, it invites us to look beyond the short term to the long-term impact of frontier AI technologies, like agentic AI, the aspirations expressed in African national AI strategies versus the actual economic reality of African states and the tradeoffs African states must make in their policy choices to be competitive in a global economy.

For Africa, this means looking beyond the hype of agentic AI to the long-term impact on economies, cultures and governance. In short, Africa’s governments must recognize that while AI offers powerful new tools, their effects will be shaped by policies and power structures – and they must act to ensure those effects serve African interests, not just global tech agendas. Africa faces a pivotal moment as AGI and “superintelligent” AI could either empower or undermine the continent. If developed with African values, rights, and priorities in mind, AGI could drive progress in education, healthcare, infrastructure, and economic growth. But if imported as foreign-controlled systems, it risks deepening surveillance, exploitation, and inequality. AGI could enhance state capacity if governments harness it for public benefit, otherwise, it may worsen inequality, concentrating wealth and innovation abroad while automating away jobs in already fragile economies. Politically, AGI also poses threats to Africa’s fragile democracies with AGI systems entrenching authoritarian control through data surveillance. Culturally, AGI systems from the West may also threaten African languages and worldviews in the AI ecosystem. Unprepared African governments could face destabilizing effects—economic shocks, social unrest, and even erosion of democracy—if powerful AGI systems begin shaping decisions beyond human control.

Africa’s AI landscape today is marked by contrasts. On one hand, millions of young Africans are digital natives, and there is growing investment in continental connectivity and data centers. Yet significant gaps persist. The continent accounts for barely 1% of global AI compute capacity, meaning most cloud, data storage and processing power is provided by foreign firms. Similarly, only about 3% of the world’s AI talent resides in Africa. These deficits are compounded by educational gaps and STEM training is uneven. This creates an AI divide where a few countries (Kenya, Nigeria, South Africa, Egypt) attract the bulk of startup funding, while many regions struggle to train AI specialists or deploy homegrown innovations. Without intervention, Africa risks remaining a supplier of raw data and a consumer of imported AI systems, rather than an innovative centre of the AI economy.

Sovereign AI and Local Agency

To address these challenges, African experts have proposed the concept of sovereign AI – the capacity for African nations to design, develop and govern AI in ways that serve local priorities. Qhala, a leading innovator of local AI technologies in Africa define Sovereign AI as a strategic framework grounded in four foundational pillars: (1) Data Sovereignty and Ethical Governance, (2) Local Infrastructure, (3) Homegrown Talent, and (4) Contextual AI Governance. In practice, this means building African datasets and cloud/cloud-like infrastructure (“own” the compute), embedding AI literacy and research in education (“educate” the workforce), and crafting governance norms rooted in community values like Ubuntu (“steward” AI with bias audits, multi-stakeholder oversight). As Qhala explains, without control over data generated on the continent, Africa cannot control its AI future. By investing in locally relevant language models (as projects like Masakhane in South Africa do) and renewable-energy data centres, African states can begin to break the cycle of dependency on external providers. The concept of AI sovereignty rests on ideas about decolonial approaches to AI governance to build back better responsible AI that centres African dignity, equity, and agency.

Data Colonialism and External Dependencies

Closely connected to the concept of AI sovereignty is data sovereignty - the sovereign control over data as a factor of production in the development of AI. While data may appear intangible on first look, the control of access to data has become a key strategic move by African states in asserting control.

Without robust local infrastructure and governance, Africa’s digital ecosystem risks replicating colonial patterns. Many scholars describe this threat as data colonialism. Large tech firms harvest African data to train AI, but the economic and epistemic value of that data often flows north. Scholars caution that these new extraction patterns mirror past colonialisms, perpetuating inequality via biased algorithms and opaque decision systems. To counteract this, African states are enacting data-governance policies and localization rules aimed at keeping data on the continent. While these laws risk other unintended consequences beyond the scope of this analysis, the real question is how do we build resilience and preparedness within African states and democracies in this period?

African AI Governance as a Decolonial Project

In July 2024 the AU adopted a Continental Artificial Intelligence Strategy that highlights data ownership, homegrown innovation, and ethical AI for development. AU officials emphasize “making AI available for socio-economic development, fostering homegrown capacity, advancing a multisectoral…governance approach and promoting innovative regulations” that protect citizens while enabling technology.

In April 2025 African Heads of State went further by endorsing the Africa Declaration on Artificial Intelligence at the Kigali AI Summit on Africa. This landmark declaration, backed by all AU member states, commits to principles of sovereignty, inclusivity and ethical development. It specifically champions data sovereignty: leaders pledged to create African-controlled open datasets and interoperable models, and to establish institutions like an African AI Council to govern them. The Declaration represents Africa’s collective insistence that AI serve the continent’s diverse linguistic, cultural, and socioeconomic contexts rather than homogenizing global norms.

The AU continental strategy reflects a broader reframing of AI governance as part of Africa’s decolonization. Effoduh and Mudeyi argue that the AU’s AI agenda can be seen as a form of technological self-determination. According to them, “the strategy acknowledges AI as a tool of global inequality, where the Global South remains a passive data provider while the Global North monopolizes AI’s economic benefits.” For them, by asserting data as a strategic resource (akin to past resource nationalizations), and by emphasizing indigenous values (community welfare, respect for local knowledge) in AI ethics, the strategy explicitly challenges the myth of technological neutrality. In practical terms, calls within the AU plan for South-South partnerships and intra-African research networks aim to reduce reliance on Northern tech giants. In this view, the authors claim that the AU strategy reimagines AI governance not just as good policy, but an act of emancipation: “African states dictate the terms of their digital future,” making the AU AI strategy itself a bold defiance of digital coloniality.

Activism, Regulation, and New Tech Visions

Building resilience for AI safety requires more than high-level strategies. It demands active civic engagement, robust regulation, and even imagining alternative paradigms of technology. It means not only “working within the system” by demanding enforcement of human rights within the AI ecosystem, but also “working against it” through activism and public pressure – and even working beyond it by cultivating entirely new tech vision. Think Tanks such as the African Observatory on Responsible AI are already sounding the alarm and are advocating for better governance frameworks that covers the entire lifecycle of AI design to deployment and adoption and training a network of stakeholders on data rights, advocating for AI safety, and lobbying governments for stronger AI oversight. At the same time, African states need to rethink the concept of sovereign AI as a matter of collective strength rather than individualised control by nation states recognising that African state economies are too small to support large scale AI ecosystems independently. This means identifying the key strategic advantages that individual countries have to support a regional AI ecosystem that can collectively advocate and restrict external pressures on practices that are harmful to the continent as a whole. This can further include adoption of public-utility models of AI infrastructure as alternatives to purely profit-driven platforms. On the policy side, harmonisation of data governance frameworks is necessary, and the AU’s Data Policy Framework sets the stage for this work to commence.

Africa stands at a crossroad. The choices made today – in policy and practice – will determine whether AI becomes another wave of external control or a genuine opportunity for African-led development. By rejecting naive technological determinism, investing in sovereign capabilities, and forging a decolonial approach to AI, African states and democracies can build the resilience needed to harness AI for the continent’s people and priorities.