This is not an argument against AI. It is an argument for sequencing. And the distinction matters enormously.
The Paradox of Premature Adoption
At a small community hospital in Gabon, the patient register is still handwritten. Nurses flip through paper ledgers, sometimes misplacing entire patient histories. Meanwhile, the global AI industry is valued in trillions. While the world debates how generative AI will transform industries, many African economies remain stuck in paper-based systems that constrain productivity, inclusion and competitiveness.
This paradox defines Africa's central challenge. The continent accounts for less than 1 percent of global data centre capacity and produces under 1 percent of global AI research output. Internet penetration stands at 38 percent, far below the global average of 68 percent. These are not mere statistics. They represent a structural gap that cannot be papered over with ambitious strategies or imported technology.
The Warning from History
History offers cautionary lessons that African policymakers would do well to heed. Ghana's rapid state-led industrialisation in the 1950s and 1960s—marked by ambitious factories and major hydroelectric investments—ultimately faltered because these industries had weak domestic linkages and depended heavily on imported inputs. The factories were built, but the ecosystem to sustain them was not.
Economist Ricardo Hausmann's work on economic complexity demonstrates that countries grow by accumulating productive capabilities—skills, institutions and interconnected sectors that enable more sophisticated production. AI can accelerate this trajectory, but only if it is layered onto economies already building those underlying capabilities. Dani Rodrik's concept of premature deindustrialisation reinforces this warning: globalisation and labour-saving technologies have narrowed the traditional industrialisation ladder, eroding the employment and capability-building benefits Africa cannot afford to lose.
Just as many countries once industrialised prematurely—importing factories before developing skilled labour, supply chains and domestic markets—Africa now risks premature automation: adopting AI technologies before building the digital foundations to harness them productively.
The Digital Dependency Cycle
The danger is not abstract. Without sequencing, Africa risks becoming the world's raw data mine—exporting information, importing algorithms and capturing little of the value. The echo with history is sobering: once raw minerals, now raw data.
This is what a digital dependency cycle looks like. AI models, platforms and governance systems are designed elsewhere, while Africa remains a consumer rather than a producer in the digital economy. The result would be a replay of extractive development—only this time, in code!
Concerns about data ownership are already materialising. Big tech companies have previously collected indigenous knowledge or language data without full disclosure, in addition to plundering the continent's resources for their requirements—a phenomenon some call techno-colonialism. There have been documented cases where biometric data was collected from communities without proper government knowledge or transparency. The issue is not data collection itself—Africa needs more data to reduce bias—but ensuring full disclosure, consent and clarity on how the data will be used.
The Job Displacement Risk
In advanced economies, AI complements aging and high-cost workforces. In Africa, it could undercut the continent's greatest asset: its young and cost-competitive labour force.
Nearly 12 million Africans enter the job market each year, but only about 3 million formal jobs are being created. If deployed hastily, AI could displace workers in key sectors—such as call centres in Kenya and Rwanda, logistics operations in South Africa, or financial back-office services in Nigeria—before alternative employment opportunities emerge. Without sequencing, automation could deepen social vulnerability and instability.
This is not a theoretical concern. It is a demographic imperative that demands careful management.
Four Priorities for Sequencing
Africa's late-mover status is not a disadvantage if used wisely. By sequencing deliberately, countries can design guardrails before diffusion accelerates—avoiding the mistakes advanced economies are now scrambling to correct.
1. Rule the data or be ruled by it. Data governance is now industrial policy. Regulatory frameworks must mandate digitisation, interoperability and data sovereignty. When governments and local firms own, analyse and control data, they shape the AI economy rather than surrender it. Gabon's directive on digitalisation, Rwanda's National Data Strategy and Ghana's Digital Economy Policy are early steps toward this sovereignty.
2. Invest in digital foundations. Digital public infrastructure—payments, digital IDs, e-signatures and local data centres—is today's equivalent of roads and power grids. In Gabon, linking small enterprises to regional payment rails expands market access while generating structured datasets essential for AI.
3. Build skills before scaling systems. Africa has the largest pool of talented human capital in the world. Training programmes are expanding, but the enabling environment and absorptive capacity must keep pace. As Nigeria's Minister of Communications, Innovation and Digital Economy, Bosun Tijani, has noted: "When we train without providing the enabling environment and absorptive capacity, these young people will not get the opportunity to participate."
4. Sequence sector by sector. Rather than broad investments, support should target high-potential sectors where Africa can build export-oriented AI-enabled services. Agriculture, health and climate resilience offer immediate opportunities where AI can solve practical problems while building local capability.
The Infrastructure Bottleneck
The hard truth is that AI requires infrastructure that much of Africa does not yet have. The continent has less than 500 megawatts of data centre capacity serving a population of roughly 1.4 billion people. By contrast, a quarter of the 150 new data centre projects in the US in 2025 exceed that capacity.
Hyperscale facilities need reliable power. In some countries, electricity penetration is less than 50 percent and that needs to increase. Energy reliability, not just price, is the binding constraint. For data centres, where investment is recovered over a long horizon, regulatory unpredictability adds another layer of risk.
A $1 billion AI data centre project in Kenya, announced by Microsoft and Abu Dhabi-based G42 in 2024, has been put on hold due to electricity requirements and financing arrangements. Kenyan officials acknowledged that the original proposal would require more power capacity than the country can currently dedicate to a single data centre. The project is a cautionary tale: computing power requires supporting infrastructure, which the continent currently struggles to provide.
What Africa Should Not Do
Africa should not rush to regulate what it has not created or does not understand. As Nthanda Maduwi of the Ntha Foundation has provocatively argued: "Job losses may be a good thing. Many of the jobs we've had have been clerical jobs—managing donor funds, writing too many reports. Maybe AI can do that and that's a good thing."
This is not an argument for complacency. It is an argument for strategic clarity. Regulation is necessary because AI is not local. When you use tools coming from outside, you need regulation—even if you don't create AI.
The Path Forward
The question is not whether Africa should adopt AI. The question is how—and when. The continent's young population is projected to grow by 450 million by 2035, offering a strong workforce for an AI-enabled economy if skills development is prioritised. AI can unlock major productivity gains in agriculture, healthcare, logistics, finance and manufacturing.
But this potential will only be realised if Africa sequences its approach: building data infrastructure, strengthening digital public goods, investing in skills and creating enabling environments before scaling AI deployment. The greatest risk is not missing the AI revolution. It is joining it too early, without the foundations to harness it productively.
The continent has seen this pattern before. It cannot afford to repeat it.
With reporting from the Brookings Institution, the African Union, the United Nations Economic Commission for Africa, the International Electrotechnical Commission, Bloomberg, GIZ, TechCabal and the OECD.
Comments (0)
No comments yet. Be the first to share your thoughts!