The Colonisation You Did Not See Coming

While Nigeria celebrated 93% AI adoption, no one asked the only question that matters: whose values are encoded in the models we are so enthusiastically feeding our minds into?

IN 1877, a British cartographer named Edward Stanford published what became the most widely distributed map of Africa in the Victorian age. It was elegant. Authoritative. Printed on thick cream paper and sold to colonial administrators, shipping companies, missionaries, and military officers across the empire. Heads of state consulted it. Treaties were drawn along its lines. Borders were negotiated using its coordinates as the final word.

There was one problem. Vast stretches of the interior were wrong.

Stanford had not commissioned explorers or interviewed African traders who knew those territories with the fluency of people who had walked them for generations. He had assembled existing European survey fragments, made educated guesses about the spaces between them, and filled in the gaps with confident cartographic authority. The mountains that did not exist on the ground existed on his map. Rivers ran the wrong direction. Kingdoms that had governed those spaces for centuries were unnamed or misnamed — their sovereignties simply erased by a man who had never visited and had no need to.

I want you to sit with that image for a moment — not as history, but as mirror. Because what Stanford did with cartography in 1877, the large language model companies are doing with human cognition in 2025. They are mapping the territory of intelligence. They are drawing the contours of what counts as knowledge, creativity, beauty, argument, and truth. They are filling in the spaces between what they can measure with confident algorithmic authority.

And the people who know those territories most intimately — whose ancestors built civilisations across them, whose proverbs encode centuries of observed wisdom, whose oral traditions carry epistemological architectures more complex than anything a Western curriculum has produced — are, once again, not holding the pen.

They are being mapped.

Now here is what makes this particular version of the story infinitely more insidious than Stanford’s map. In 1877, Africans did not choose to be mapped. The cartographers came regardless. But in 2025, we are handing over our maps voluntarily. We are uploading our intelligence, our creative output, our linguistic patterns, our problem-solving architectures — into systems specifically designed to learn from them, extract patterns from them, and use those patterns to serve markets we do not control.

We are not the territory being mapped this time. We are the cartographers handing over the instruments.

I. THE DISRUPTIVE HOOK — THE CELEBRATION THAT BECAME A CONFESSION

In 2024, a Google and Ipsos study confirmed what many in Nigeria’s tech community had long suspected: Nigeria leads the world in AI adoption, with 93% of internet users engaging with AI tools regularly. The announcement rippled through the ecosystem with the energy of a national achievement unlocked.

Conferences were themed around it. Newspapers ran headlines positioning Nigeria as the unlikely vanguard of the AI revolution. Policymakers cited the figure as evidence that Africa was not being left behind — that the continent, this time, was at the frontier.

Nobody paused to ask what, precisely, had been adopted.

Here is what that 93% figure actually means when you look at it closely. It means that the majority of Nigeria’s internet-connected population is regularly feeding its intellectual output — its writing, its thinking, its decisions, its creative expression — into systems whose training data is estimated to be over 85% English-language content dominated by Western cultural contexts.

It means Nigerian professionals are using tools that have learned to see the world through a lens that did not include their grandmothers’ proverbs, their community’s conflict resolution architecture, their language’s particular way of encoding respect through tonal register, or their culture’s understanding of time as circular rather than linear.

It means the most AI-adopted nation on earth is also, in a very specific and measurable sense, the nation most exposed to what I will name plainly: Algorithmic Cultural Displacement.

The colonisation of the 21st century does not arrive by ship. It does not require soldiers, administrators, or even a policy. It requires only a subscription — and the slow, invisible erosion of the cognitive infrastructure by which a people understands itself.

II. THE INVISIBLE ROT — WHAT GETS LOST WHEN THE MAP IS WRONG

There is a phenomenon in linguistics called language attrition. It describes what happens to a person’s native tongue when they are immersed, over time, in a second language that holds higher social and economic status. The mother tongue does not disappear overnight. It erodes. Vocabulary thins first. Then syntactic structures begin to simplify. Then the conceptual categories that only exist in the native language — the ones with no direct translation, the ones that encode a worldview — begin to recede. The person still speaks their language. But they speak a lighter version of it. A version that has quietly surrendered its most complex registers.

Cognitive sovereignty attrition works identically. It does not announce itself. It does not arrive with an ultimatum. It operates through repeated, accumulated deference to a system that holds higher operational status — a system faster, more confident, and more instantly rewarding than the process of thinking things through from one’s own cultural and moral foundations.

I am watching it happen in real time. In boardrooms and ministries, in creative studios and university departments, in the WhatsApp groups of Nigeria’s professional class. The tell is not that people are using AI. The tell is how they are using it — as the first resort rather than the last. As the arbiter of what sounds authoritative rather than a tool one interrogates with one’s own authority intact.

The displacement operates across three registers simultaneously, and it is the convergence of all three that makes it structurally dangerous rather than merely inconvenient.

The first register is aesthetic. When African creative professionals use AI tools trained predominantly on Western creative output, those tools carry embedded aesthetic hierarchies — what a ‘good’ story structure looks like, what a ‘sophisticated’ argument sounds like, what a ‘professional’ tone feels like. These are not neutral standards. They are culturally specific preferences that have been elevated to the status of universal norms through the sheer weight of the training data. African creators do not feel the pressure directly. They feel it as a faint, persistent sense that their natural creative instincts need to be refined — elevated — toward a standard that they absorb from the tool without ever consciously agreeing to.

The second register is epistemological. African intellectual traditions carry specific ways of knowing — communal verification, oral argumentation, the authority of lived experience over documented evidence, the legitimacy of wisdom that cannot cite a peer-reviewed source. These epistemologies are not primitive. They are sophisticated alternative architectures for establishing truth. But they are invisible to AI systems trained on Western academic and professional text. When African professionals use those systems to think, they are operating in an intellectual environment that systematically undervalues the cognitive tools their culture gave them.

The third register — the most dangerous — is psychological. When you repeatedly produce your best thinking through a system that remixes it back to you in a style that feels more polished, more authoritative, more universally legible than your own unmediated voice, something shifts in your relationship to your own intelligence. You begin to distrust the unpolished version. You begin to hear your own unmediated thinking as somehow insufficient. And in that gap — between your raw intellectual authority and the AI-mediated version of it — the colonisation does its deepest work.

I want to name something here that almost no one in this conversation is willing to say directly. The companies building these AI systems are not malicious. They are not conspiring against African intelligence. They are doing something far more mundane — and far more dangerous. They are building products for the markets that have the most purchasing power and the most data. African voices are underrepresented in their training data not because of hostility but because of economics. And the effect is the same regardless of the intent.

The map is wrong. It is being used. And the territories it misrepresents are being governed by its errors.

I want to introduce the second foundational doctrine of this column — one that will reframe every conversation you have about AI adoption from this point forward.

The Doctrine of Algorithmic Cultural Sovereignty.

It states this: Every AI tool carries a cultural fingerprint. The fingerprint is not visible in the product documentation. It is not disclosed in the terms of service. But it is present in every output the tool produces — in the aesthetic preferences it elevates, the argumentative structures it rewards, the conceptual categories it recognises, and the kinds of intelligence it is constitutionally incapable of representing.

Algorithmic Cultural Sovereignty is the capacity of an individual, institution, or nation to understand that fingerprint — to interrogate it, to refuse its hidden authority, and to ensure that the cultural intelligence embedded in one’s own traditions, languages, and epistemologies remains the governing architecture of one’s creative and intellectual output regardless of the tools being used.

It is the discipline of ensuring that when you use a foreign map, you still know where you actually are.

This doctrine has three immediate strategic implications. The first is for individual professionals: your native cognitive architecture — the way your culture trained you to think, to argue, to create, to resolve ambiguity — is not a liability to be smoothed over by AI. It is your primary competitive asset in an age when every AI-mediated output converges toward the same aesthetic mean. Your differentiation lives in what the tool cannot reach.

The second is for institutional leaders: every AI adoption decision your organisation makes is a cultural governance decision. The question is not whether the tool works. The question is what the tool does to your organisation’s ability to think with its own intelligence, in its own cultural register, toward its own strategic ends. The organisation that cannot ask that question has already answered it by default.

The third is for governments and policymakers: the Nigeria Data Protection Commission and the frameworks being developed around African data sovereignty are necessary but not sufficient. Data protection governs what happens to your information. Cultural sovereignty governance must address what happens to your cognition — how AI systems shape the way your population thinks, creates, and decides at scale. That dimension is almost entirely absent from current African AI policy, and its absence is not an oversight. It is the most consequential governance gap on the continent.

Power & Sovereignty: The Infrastructure of Intelligence

There is a reason the most powerful empires in history did not merely conquer territory. They controlled the infrastructure through which territory was understood — the maps, the legal codes, the educational curricula, the languages of commerce and governance. The control of interpretive infrastructure is always more durable than the control of land, because it operates inside the minds of the governed rather than merely upon their bodies.

The large language model is the interpretive infrastructure of the 21st century. It is the system through which knowledge is generated, mediated, and legitimised at scale. And its ownership is concentrated in a handful of organisations headquartered in two countries, trained on data that reflects the cultural assumptions of a specific segment of global humanity, and governed by regulatory frameworks designed to serve the interests of their home markets.

Let me be direct: this is a power story, not a technology story. And Africa is, once again, in a position where the infrastructure of interpretation is being built without it — and being adopted by it at scale.

Behavioural Economics: The Compounding Cost of Borrowed Thinking

Economists speak of path dependency — the phenomenon by which early choices constrain future options, not because the alternatives are unavailable, but because the accumulated weight of previous investment makes switching costly. A nation that built its entire rail network on a specific gauge finds it prohibitively expensive to change, even when a better gauge becomes available. The infrastructure is locked in.

AI adoption creates cognitive path dependency. The more an individual, institution, or nation relies on a specific AI ecosystem to mediate its intellectual output, the more its cognitive infrastructure shapes itself around that ecosystem’s assumptions. Thought patterns reorganise. Aesthetic preferences realign. Institutional knowledge is increasingly expressed in the categories that the AI tool recognises and rewards. The cost of switching — or of maintaining genuine cognitive independence — rises with every month of continued adoption.

Nigeria is, right now, in the early stages of the most consequential cognitive path dependency event in its history. The decisions made in the next three years — about how AI is adopted, governed, and integrated into professional, educational, and governmental practice — will shape the cognitive infrastructure of the nation for a generation.

Algorithmic Governance: The Policy Gap That Costs a Continent

My observation of behaviourally-adaptive AI governance across multiple jurisdictions reveals something that should be stated plainly: there is currently no African government with a comprehensive AI cultural sovereignty framework. There are data protection laws. There are digital economy strategies. There are AI ethics committees in various stages of formation. But the specific question — how do we ensure that AI adoption does not systematically erode the cognitive and creative sovereignty of our population — is not on any policy agenda I can identify.

The Nigeria Data Protection Commission’s work is important. But data protection, at its core, concerns what happens to information once it is collected. Cultural sovereignty governance must go deeper: it must address what happens to the human beings who are generating that information, and what happens to their capacity for original, culturally-rooted thought when the tools they use most are constitutionally incapable of recognising the deepest registers of their intelligence.

Stanford’s map governed territory for a hundred years because no one with authority insisted on using a different map. The question for African governments is not whether they will have an AI policy. The question is whether their AI policy will include the right map.

The framework below is not a checklist. It is a diagnostic instrument built from close to thirty years of direct observation across boardrooms, ministries, creative institutions, and executive coaching engagements. It measures the degree to which an individual or institution is operating with genuine cognitive and cultural sovereignty in an AI-saturated environment — or being mapped by the system without knowing it.

THE IGNORED PATH: A Continent Mapped by Others — Again

By 2035, the large language model has become the primary interface through which Nigeria’s professional class accesses knowledge, produces creative output, and makes decisions. The models in widest use are updated versions of systems trained in 2024 and 2025 — the critical window when African intellectual output was most abundant but African governance of its use was most absent.

The aesthetic output of Nollywood converges, over that decade, toward the narrative structures that Western distribution algorithms reward — three-act arcs, individualised protagonists, linear resolution. The richly communal, morally complex, time-fluid storytelling architecture of Yoruba and Igbo narrative tradition becomes ‘local flavour’ within a globally legible template — present in surface detail, absent at the structural level.

Nigerian professionals speak with authority. But the authority sounds — in its deepest registers — like the authority of a system that was not built for them. The map is in use. The territory it misrepresents does not appear on any policy agenda. Because by 2035, no one remembers what the territory looked like before the map became the reality.

THE ADOPTED PATH: A Continent That Wrote Its Own Map

The alternative begins with a decision — not a policy, not a summit, not a strategy document. It begins with individuals and institutions deciding, one by one and institution by institution, that their cognitive and cultural sovereignty is not a soft value. It is a hard strategic asset.

African-trained AI models, built on indigenous language corpora and tested against African epistemological standards, become the foundational infrastructure for the continent’s most important creative and governance work by 2032. The organisations that invested in building them are not philanthropists. They are the most powerful media companies, technology firms, and policy institutions on the continent — because they own the interpretive infrastructure that everyone else depends on.

Nigerian and broader African creative industries — operating from a position of Algorithmic Cultural Sovereignty — produce the most distinctive, least replicable creative output in the global marketplace. Not despite the AI revolution. Because of their disciplined refusal to let it define the terms of their imagination.

Edward Stanford’s map of 1877 was not the last wrong map of Africa. In 1884 and 1885, at the Berlin Conference, the European powers sat around a table — no African representatives present — and divided the continent along lines that bore almost no relationship to the actual political, cultural, and ethnic architectures of the people who lived there. Those lines are still bleeding.

The borders they drew were not just geographic errors. They were cognitive impositions. They told the continent what shape it was. And once the map was on enough desks, in enough government buildings, used in enough treaties — the map became harder to challenge than the territory. Because the territory could be walked. The map was everywhere.

I am asking you to understand that the large language model is a map. It is the most sophisticated, most widely adopted, most deeply persuasive map of human intelligence ever constructed. And it is being drawn, right now, without adequate African input, without adequate African governance, and without adequate African awareness of what is being encoded in its training data while the window is still open.

The window closes not with a proclamation but with a threshold — the moment when enough decisions, enough institutional workflows, enough cognitive habits have reorganised themselves around the map’s authority that the cost of insisting on the territory’s actual shape becomes prohibitive.

We are not at that threshold yet. That is the only reason this column matters — not as analysis, but as alarm.

The leaders who act on this — who build cultural sovereignty into their personal discipline, their institutional governance, and their national policy agendas — will not merely survive the AI century. They will be the ones who finally hold the pen.

3 VIRAL HEADLINE VARIATIONS

1. The Last Time Someone Else Drew Africa’s Map, the Borders Are Still Bleeding. The AI Era Is Drawing a New One.
2. Nigeria Leads the World in AI Adoption. It Also Leads the World in Voluntary Cognitive Displacement. These Are Not Unrelated Facts.
3. In 1877, a British Cartographer Mapped a Continent He Had Never Visited. In 2025, Silicon Valley Is Doing the Same Thing to African Intelligence.

2 HIGH-TENSION LINKEDIN HOOKS