Why Africa Must Move from Using Artificial Intelligence to Curating the Intelligence of Its Own Existence
By Niabari Fakae
The next great question about artificial intelligence is not simply what machines will be able to do. It is whose knowledge will teach them, whose realities will shape their answers, whose languages they will understand, and who will own the value created when human experience becomes machine-readable.
For Africa, this is not a distant philosophical question. It is an economic, cultural and strategic question unfolding now. The continent is using AI tools in education, business, communication, entertainment, finance and public life. But using systems built elsewhere is only one part of the story. A more consequential question is whether Africa will deliberately curate the knowledge, data, context and intellectual resources that help shape the systems the world will increasingly depend on.
The distinction matters. A continent can become an enthusiastic consumer of technology while remaining underrepresented in the knowledge systems that power it. It can supply raw material without controlling the standards by which that material is organised, interpreted, licensed or converted into economic value. It can be visible as a market and still remain faint in the datasets through which its languages, histories, creative expressions and social realities are represented.
The missing layer is not merely data. It is curation.
Data does not become intelligence simply because it has been collected. It must be described, contextualised, checked, connected and governed. A proverb detached from its language and social setting may be reduced to a literal sentence. An oral history without provenance may lose the relationship between memory, place and the person who carries it. A song, pattern, photograph or craft record without permissions and attribution may become available to a system while the people who created or preserved it remain invisible.
Curation is the work of making knowledge usable without pretending that context does not matter. It asks what a record means, who has authority to share it, what may be made public, how it should be labelled, what is uncertain, who can correct it and what conditions should govern its future use. In an AI economy, these are not administrative details. They influence what systems can recognise, what they overlook and what they reproduce incorrectly.
UNESCO has described African languages as a blind spot in AI, noting that systems trained on dominant-language material can struggle with languages such as Hausa or Zulu. The African Union’s Continental AI Strategy also identifies gaps in the quality, inclusiveness and availability of locally produced datasets. These concerns point to a larger task: building knowledge infrastructure that reflects African realities and allows African people to participate in shaping AI.
From AI Adoption to African Intelligence Infrastructure
I have been thinking about this transition through the emerging idea of African Superintelligence.
The phrase is deliberately ambitious, but it needs to be understood carefully. It is not a claim that Africa can create machine superintelligence merely by collecting local data. Nor does it suggest that Africa possesses one unified intelligence or a single cultural identity.
Africa is plural: thousands of languages, distinct communities, layered histories, knowledge traditions, creative economies and contemporary experiences.
African Superintelligence, as a research proposition, is the collective capacity to curate, connect, protect and apply this diversity of knowledge through people, institutions, data systems and AI.
It is an infrastructure of contribution. It asks how knowledge that has existed in homes, markets, workshops, archives, farms, waterways, studios and oral traditions can be responsibly documented and made useful for education, research, innovation and economic activity, without treating every form of knowledge as material that should be extracted or opened to everyone.
This requires a change in the questions we ask.
Instead of asking only which AI tool can make a task faster, we must ask which African realities the tool can understand. Instead of asking only how businesses can use AI, we must ask how African businesses, researchers, creators and communities can contribute to the data, methods, evaluation and governance of AI. Instead of treating local context as an obstacle to be translated away, we must treat it as knowledge that can improve the relevance and usefulness of systems.
Why This Is an Economic Question
The AI value chain extends beyond the final application. It includes data collection, annotation, language resources, software, model adaptation, computing infrastructure, evaluation, deployment, maintenance, intellectual property, licensing and services.
If African institutions participate only at the point of consumption, much of the value created upstream and downstream may accrue elsewhere.
If they develop trusted data resources, specialised services, local models, evaluation tools and partnerships, they can create additional roles for researchers, linguists, archivists, educators, artists, developers, entrepreneurs and knowledge holders.
That possibility should not be romanticised.
Data projects cost money. High-quality curation takes time. Compute and connectivity can be expensive. Markets may not emerge automatically, and not every dataset should be commercialised.
Some resources may be best maintained as public-interest infrastructure. Others may support cooperative licensing, research partnerships, creator-led businesses or carefully governed commercial services.
The point is to design for fair participation from the beginning rather than assume economic benefit will arrive after the data has been taken.
This is where the idea connects to the creative and orange economy.
African creativity is not limited to finished artworks or entertainment products. It includes languages, symbols, techniques, histories, rhythms, performance practices, design traditions and social meanings.
Properly governed cultural data can support discovery, preservation, education and new creative services. But creators and communities must have meaningful choices over consent, attribution, licensing and benefit-sharing. Cultural richness must not become an excuse for unlicensed extraction.
Reimagine Africa: From Conversation to Infrastructure
The wider conversation around Reimagine Africa, alongside discussions about AI in the emerging world, offers a useful intellectual setting for this question.
Reimagining Africa cannot end with the language of possibility. It must also ask what structures, capabilities and shared commitments make that possibility real.
In the context of AI, that means moving from celebrating African creativity to investing in systems that preserve, describe, connect and responsibly activate African knowledge.
The essential shift is from seeing Africa primarily as a recipient of imported innovation to recognising its capacity to contribute to the technologies, datasets, ideas and institutions that shape the future.
The work ahead is to translate that ambition into practical commitments: shared research, locally grounded datasets, community-led governance, technical skills and sustainable institutions.
The opportunity is to make the conversation practical.
What would it mean for universities to work with language communities to create consent-based corpora?
What would it mean for archives and creative organisations to describe their collections so they are discoverable without surrendering rights?
What would it mean for local researchers to test AI systems against African languages and contexts, document failure and publish useful evaluation resources?
What would it mean for a creator or community to know when their material is used, under what terms, and how value returns to them?
These are not abstract questions. They are the beginnings of an infrastructure agenda.
The Ethics of Collecting What Has Been Overlooked
A serious African intelligence agenda cannot be built on the idea that everything should be collected in bulk.
Some knowledge is private, sacred, restricted, commercially sensitive or held under community obligations.
Some people may not want their voices, images, stories or creative work used to train a model. A dataset can reproduce bias as easily as address it. A system that performs well on a benchmark may still misunderstand a community or produce damaging interpretations.
The ambition, therefore, must be intentional curation rather than indiscriminate accumulation.
Communities need a say in what is recorded and what is excluded. Contributors need clear information about use, access, retention, licensing and withdrawal where feasible.
Data should carry provenance and context. Evaluation should include people who understand the language and knowledge domain. Governance should address security, privacy, intellectual property, commercial reuse and fair benefit-sharing.
The measure of success is not the size of the archive alone, but whether it is trusted, useful, sustainable and legitimately governed.
A Doctoral Question for a Changing World
These questions motivate a proposed doctoral research programme on the transition from AI use to the curation of African intelligence.
The research would examine how selected African knowledge can be curated into AI-ready resources, how those resources affect defined machine-learning tasks, and what governance and economic arrangements allow African contributors to retain agency and capture value.
It could begin with a bounded pilot in language, cultural expression or another locally chosen knowledge domain, rather than attempting to represent the entire continent in one project.
Such research would bring culture and media into conversation with computer science, information studies, data governance and economic development.
It would treat African intelligence not as a slogan but as a set of questions that can be investigated: What is missing? Who decides? What does responsible curation require? Can the resulting data improve an AI system in measurable ways? What kind of institution or market can sustain the work? And who benefits?
The Future Should Not Only Know Africa. It Should Learn with Africa.
The global AI conversation is also a conversation about memory, representation and power.
The information that survives into digital systems can influence what future generations learn, what businesses discover, what researchers can investigate and what technologies consider possible.
If African knowledge remains fragmented or absent, the world’s increasingly automated systems will continue to offer incomplete accounts of human experience.
Africa’s response should be neither withdrawal nor uncritical adoption.
It should be deliberate participation: building the skills, partnerships, standards, infrastructure and governance that allow African knowledge to contribute on fair terms. That includes machine learning and computing, but also archives, language work, creative practice, law, ethics, community stewardship and economic design.
The opportunity is larger than teaching people to use AI.
It is to create the conditions under which African knowledge can shape what AI becomes, while ensuring that the people who carry that knowledge are recognised and respected.
The next phase is not simply to ask how Africa can catch up with the intelligence economy. It is to ask how Africa can help define the intelligence that economy values.
That is the question worthy of study: how we move from using artificial intelligence to intentionally curating the intelligence of our own existence, and how that contribution can strengthen Africa’s intellectual agency, economic participation and place in the world that is coming.
Niabari Fakae
References and Further Reading
- African Union, Continental Artificial Intelligence Strategy (2024): https://au.int/en/documents/20240809/continental-artificial-intelligence-strategy
- African Union, Adoption of the Continental AI Strategy and African Digital Compact (2024): https://www.au.int/en/pressreleases/20240617/african-ministers-adopt-landmark-continental-artificial-intelligence-strategy
- UNESCO, African Languages: The Blind Spot of AI (2026): https://www.unesco.org/en/articles/african-languages-blind-spot-ai
- UNESCO, Project on African-Language Datasets for AI: https://core.unesco.org/en/project/A210632081