I was seven years old, brushing my teeth in the bathroom, when for the first time in my life I heard a gunshot.
In the days before, there had been signs I was too young to read properly. My father in front of the radio more than usual. Fragments of adult conversation I’d catch and not understand. A tension in the house that I couldn’t have named at seven but could still feel, the way children always feel what adults are trying to hold back from them.
It was 1998, somewhere between July and August, school holidays, in the DRC. The morning had started normally. Then came the rattle of an AK-47, and something in me went still. From the top of my head to my toenails, I was genuinely paralyzed. My sisters called me to join them in the room where everyone was gathering. My mother was abroad getting medical treatment, so it was just my father, four of my siblings, one of my mother’s little brother and me, the youngest. I was trying to understand something my seven-year-old mind had no frame for.
It didn’t stay at AK-47s. That day and the following days the sound had escalated to something heavier, weapons of mass destruction shaking our house every few minutes, the walls literally moving around us. With time, the sound became almost like a lullaby. You’d think constant gunfire would keep a child awake with terror every single night ? Think again. My body learned to sleep through it. The fear never actually left me, it just stopped being loud enough to keep me up and shaking like in the first days. It’s strange and a little frightening, how quickly chaos can start to feel like the norm once it repeats enough times.
We spent close to two months like that. No water, no electricity. My sisters walked to the nearest river I did not even know existed until then, to fetch water, and every time they left, we sat with every rumor and story we’d heard about people who went out and didn’t come back. None of us ever really knew if they’d walk through that door again. We went to bed without knowing if we’d wake up. At night we laid our mattresses down in the corridor that led to the bedrooms, my father closest to the front door, me at the very back, in case something happened. I remember him telling us we might need to move to a more secure spot, because the fighting was getting closer.
What made those moments feel survivable, I think, was watching the adults around me. As long as the fighting stayed a few kilometers away rather than on our street, people kept moving, checking on neighbors, going about fragments of ordinary life, refusing to fully stop. I didn’t see what people in the eastern DRC were seeing at the time, the worst of it. What I lived was smaller than the worst but it was enough. Enough to show me, at seven years old, both sides of community at once: the side that fractures under pressure, and the side that keeps checking on the neighbor anyway, even while the ground is shaking.
I didn’t have words for any of that at seven. Honestly, I’m only finding the words now, decades later, in the middle of this AI era, writing about it so publicly for the first time in my life, trying to explain why I hold African communal philosophy in this AI era the way I do. But the process of understanding it started that year, at that age, whether I had the language for it or not.
That experience hasn’t fully let go of me. To this day, any gunshot like sound, a car backfiring, a firework, a sound I logically know isn’t a weapon sometimes sends me straight back to that morning. Not as a memory I recall, but as a feeling I re-enter. The same paralysis, head to toenails. The same seven-year-old girl standing at the sink not understanding what’s happening to her world. It’s instant, and it’s exact, and it doesn’t ask my permission. I’m telling you this not for sympathy, but because I think it matters for what comes next in this piece. When I write about community breaking and holding at the same time, I’m not describing a concept I encountered in a book. I’m describing something I’ve lived with most of my life.
Before 1998, I couldn’t have told you life was perfect. But I had never felt what it was to have community break. That chaos was the first time I understood otherwise and, strangely, the first time I understood that it doesn’t fully break either. It bends. Some of it holds. I won’t get into why that war happened(it’s well documented elsewhere) but I’ll say this plainly: that was Africa breaking Africa.
That early experience is foundational to why I approach Ubuntu (the broader idea of African communal philosophy) with a healthy dose of caution. Not because I don’t believe in it. I lived enough of it to know it’s real, and I’ve argued before, in my own past writing, that Ubuntu is a genuinely African philosophical contribution, one worth putting directly against the individualism baked into so many AI systems coming out of the West. I still believe that and I’m not against Ubuntu.
But I’ve also lived through community fracture, and I’ve watched us do that fracturing to each other. So when people talk about community as if it’s simply, uncomplicatedly Africa’s nature, a gift we’re offering the world, no cracks in it, I have to pause because it’s more complicated than that. And I should say clearly: Ubuntu is one African philosophy among many, not the only one, and not something that speaks for the whole continent on its own. Treating it as if it does is its own kind of oversimplification, even when the intentions behind that are good. I think the complicated, fuller version matters, especially now, as this same idea gets used to shape how AI is supposed to work on this continent.
Community is not an African invention. It’s a human default and 1998 wasn’t a one-off. I’ve watched this same pattern (community breaking and holding at once) play out again since, closer to where I live now. Take xenophobic violence in South Africa: attacks on foreign nationals, often other Africans, by people who are themselves living under enormous economic pressure. That violence is itself a symptom of community breaking down, precarity, scarcity, and fear turning neighbor against neighbor, the same corrosion I saw as a child, just wearing a different face. And yet, even inside those events, you’ll find people sheltering the very people being targeted, communities organizing to protect strangers, ordinary individuals refusing to let the fracture become total. Ubuntu breaking and Ubuntu holding, in the same street, on the same day.
I bring this up because it isn’t only an African pattern. You’ll find the same double movement anywhere community comes under pressure, in the way small towns pull together after a disaster even as the same pressures that caused it expose old resentments, in the way any society, anywhere, tends to fracture and cohere simultaneously when it’s under threat. That, to me, is the real evidence for what I actually believe: this isn’t Africa’s special trait. It’s the human default showing up under stress, everywhere it’s tested. What differs is not whether the instinct exists, but how completely it’s been dismantled by the systems built on top of it.
No human being becomes a person in isolation. Every human infant survives because someone feeds them, teaches them, protects them. Every human society, everywhere on earth, organized itself around kinship, obligation, and mutual care before it organized itself around anything else. Community isn’t a cultural preference some regions picked and others didn’t. It’s closer to a biological and social precondition, you can’t have humans without it.
So if that’s true everywhere, why does “communalism” get filed under African philosophy, while individualism gets filed under just how the world works?
Community doesn’t erode on its own. It often gets broken by conquest that gets turned into a system, not just conquest as a single event. By a group deciding to win at another group’s expense, and then rebuilding the whole economy around that win rather than letting it stay a one-time disruption.
That distinction matters, because conquest by itself doesn’t automatically produce individualism. Plenty of societies expanded through war and stayed deeply communal, as I’ll get to. What seems to matter is whether land and survival get pulled away from the many and concentrated in the few, people separated from their own means of survival and made dependent on wages rather than kin. Do that once, it’s an event a community can absorb. Do it repeatedly, generation after generation, until it becomes the actual operating system of an economy and a government, that’s when a society stops running on community and starts running on conquest.
This is the part that matters: that happened in Europe first, to Europeans. Long before colonial ships left European ports, European peasant communities were already being broken apart, common land fenced off and privatized, small farmers pushed off land they’d worked communally for generations, people turned into wage laborers dependent on employers rather than kin. I don’t think this is the only explanation for how individualism became Europe’s dominant operating system; the Reformation’s emphasis on individual conscience, Enlightenment humanism, the growth of cities all played their part too, and serious historians would weight these differently than I do here. And “the West” was never one thing either, mutual aid networks, labor movements, the commons before enclosure, still carried communal instincts that never fully vanished. What I’m describing is which logic won and got built into institutions and exported globally, not a total erasure of every communal impulse there. But I do think internal conquest (economic conquest, of one class over another) was a significant and often underweighted part of that story, hollowing out community from the inside well before it went looking for conquest abroad.
Once that internal transformation was complete, the same logic scaled outward. Colonialism wasn’t a philosophy arriving in Africa, it was a system that had already dismantled community at home now doing the same thing elsewhere, this time across oceans and borders, with the additional device of race to organize who got extracted from.
Yes. Precolonial African history has plenty of it. Kingdoms that expanded through warfare, empires built on tribute, rivalries and subjugation between groups. Anyone who says the continent was a single unbroken circle of communal harmony before Europeans arrived isn’t telling the truth, and that kind of romanticizing doesn’t help anyone.
But there’s a difference between conquest happening and conquest becoming the total operating system of a society. In much of precolonial Africa, expansion and rivalry coexisted with community-based structures, a kingdom could grow through war and still run internally on kinship obligation, tribute-sharing, and reciprocity. This is exactly the proof for the point I made earlier: conquest on its own doesn’t automatically dissolve community. Conquest was an event these societies absorbed and metabolized within a broader system that was still, at its core, built on relationship and obligation.
What colonialism did differently was make conquest the entire architecture, permanent extraction as the default relationship, not an episodic disruption inside a larger communal order. That’s the actual difference. Not “Africans are communal, Europeans are individualist.” It’s: in one place, conquest-logic took over completely and became the system; in the other, it kept colliding with a communal substrate that never fully gave way partly because it was interrupted by the very colonialism it’s now praised for surviving. This is the actual reason Ubuntu still feels so visible, so evident, on this continent: not because the instinct is uniquely ours, but because the process that usually buries it under a system never got the chance to finish here.
Here’s where this stops being a history lesson and becomes something with real stakes today. Watch how AI governance actually happens on this continent. A new AI system launches somewhere in Silicon Valley or Beijing. It causes some kind of harm, or a scandal, or a controversy and Europe or the US drafts a law in response. Months or years later, African countries start drafting their own AI policies, and those policies borrow heavily from the ones already written elsewhere: similar categories of risk, similar definitions of harm, similar assumptions about what a “responsible AI ecosystem” should look like. This is the pattern. Almost every AI governance document coming out of the continent right now leans on frameworks built somewhere else, for problems shaped by conditions that don’t fully match ours.
That is called reactive governance, and there are two layers to it. The first layer is simple: reacting to a crisis after it happens, instead of anticipating your own. The second layer is subtler and more damaging: even the tools you reach for to respond to that crisis were designed elsewhere, for a different kind of society, with different histories of harm. So you’re not just late to the problem, you’re solving it with someone else’s toolbox, built for someone else’s house.
This is where my personal story actually connects to something structural. If your institutions spent generations absorbing shocks that were imposed on you from outside (colonial administration, extraction, borders and systems designed for someone else’s benefit) you end up organizationally built to respond, not to originate. That’s not a lack of ambition or intelligence. It’s what happens to the architecture of a place when, for a very long stretch of its modern history, it wasn’t the one deciding what happens to it. A crisis-response playbook built for a Silicon Valley company managing a PR problem is not the same tool as one built for a society that has spent generations absorbing shocks it didn’t choose. Using the first tool for the second job is exactly what “importing frameworks” looks like in practice.
And here’s where Ubuntu gets pulled back in, in a way that should make us uncomfortable. Because communalism has become the go-to language for “responsible AI, African style,” it’s easy for that language to get used as decoration rather than substance. A company can say its AI approach is “guided by Ubuntu values” while the actual data labeling and content moderation work powering its systems (often done by African workers, for low pay, under difficult conditions) has nothing communal about it at all. The philosophy gets quoted; the extraction continues underneath it. That’s not African communalism failing. That’s African communalism being used as a costume for the exact same conquest-logic we’ve been talking about this whole time, just wearing better language.
And you’d notice that even here, in the AI space itself, the same breaking-and-holding pattern keeps playing out. We write strategies, policy papers, ethical frameworks, all invoking Ubuntu by name. That’s the language. But where’s the practice? Where are the AI systems actually built on reciprocity instead of extraction, actually returning value to the communities whose data and labor made them possible, actually governed by the people they’re meant to serve rather than governed for them from somewhere else?
I went looking for the answer while writing this. What I found was a steady, growing stream of papers, frameworks, and scorecards proposing Ubuntu-based AI governance, several from just this year alone, including a proposal for an “Ubuntu AI Framework” and Scorecard, academic models for “Ubuntu-aligned” data governance and labour platforms, and South Africa and the AU successfully getting ubuntu named in a G20 leaders’ declaration. That’s the breaking side: real effort, real thinking, but still mostly language.
What I found on the practice side was, essentially, two names, not the abundance you’d hope for. The one that kept resurfacing across every search, every year, going back to 2018, was Masakhane, a grassroots initiative building open-source natural language processing tools for African languages, run collaboratively by the researchers and communities who actually use what it builds, its own name means “we build together” in isiZulu, and its founding philosophy is drawn directly from Ubuntu. It’s real, it’s lasted years, and it’s grown into a network of thousands of contributors across the continent. The other is UbuntuGuard, a newer AI safety benchmark built from real expert input across several African countries, testing whether AI systems actually comply with locally defined policies rather than imported ones. Both are operating work. But two against a fast-growing shelf of papers and frameworks, is still the point.
Both are genuine, operating work, not proposals. Two sustained practical examples against a fast-growing shelf of papers and frameworks is not a scientific ratio, and I’d genuinely welcome being corrected with more names. But when a search for “Ubuntu and AI” turns up far more essays about the philosophy than institutions actually practicing it, that gap is the point I’m making. If the answer to “where’s the practice?” is mostly documents and speeches, that’s the breaking. If there are real and fully operating, lived examples (a data cooperative, a community-owned model, a governance process that genuinely includes the people affected) that’s the holding. Right now, we have far more of the first than the second, and until that changes, invoking Ubuntu in an AI strategy paper is closer to a citation than a practice.
So when I say Ubuntu shouldn’t be treated as our automatic answer to AI ethics, this is what I mean in practice: name the philosophy honestly, don’t let it be borrowed as branding, and build governance that responds to what’s actually happening on this continent (including the labor conditions, the data infrastructure, the ownership questions) instead of frameworks copied from elsewhere and dressed up in communal language to sound more relatable.
Not that Africa is uniquely communal and the West is uniquely individualistic. That framing flatters everyone into a story that isn’t true and makes the real explanation invisible.
What I’m arguing is that community is the human default everywhere. Individualism is what emerges when conquest (internal or external) becomes a system rather than an event. Africa isn’t more “naturally” communal than anywhere else. It’s simply a place where that conquest-into-system transition got interrupted, delayed, and imposed from outside rather than organically generated from within and possibly scaling outward the same way it happened elsewhere, which is also exactly why African institutions now find themselves reacting to governance frameworks instead of authoring their own.
Universal doesn’t mean identical, though, and I want to be careful not to trade one flattening for another. Saying community is a human instinct doesn’t erase what’s specific about Ubuntu, how it defines personhood, obligation, and what it means to become fully human through others, in ways that aren’t the same as Western communitarian thought or other relational philosophies elsewhere. The instinct may be universal. The particular shape Ubuntu gives it is still its own, still worth taking seriously on its own philosophical terms, not just as a data point in an argument like this one.
Ubuntu isn’t a cultural export we should be proud to offer the world as our special contribution. It’s what’s left of something that used to belong to everyone, in a place where the machine that usually erases it didn’t finish the job yet.
And that "yet" is the part I actually want to leave you with, not as a prediction, but as a warning I'd like to be wrong about. Colonialism didn't finish dismantling community here. It didn't have the tools, or the time, or the total reach. What it started, but couldn't complete, AI has a real shot at finishing, quietly, efficiently, without needing colonial administrators on the ground. When AI systems extract African resources, data and labor while running on a logic that treats people as inputs rather than as community, when governance keeps arriving late and borrowed, when "Ubuntu" gets used as a marketing word instead of a lived practice, that's the same machine, just running on code instead of colonial administration. Different uniform, same job.
This is where I land, and why I opened with my own story instead of a definition. My approach to Ubuntu isn’t reverence and it isn’t rejection. It’s the grain of salt that comes from having lived both halves of it, watched it break at seven, watched it hold in the same breath, watched it break again since, in different rooms, for different reasons. That’s the only honest place I can argue from. And from that place, what worries me isn’t whether Ubuntu is real. It is. What worries me is what’s left of it on this continent (the actual, lived, unfinished remainder) quietly turning into a word put in a slide deck while the extraction underneath stays exactly the same. And I don’t just mean AI companies doing this to us. I mean the African AI strategies, the essays, the policy papers, invoking Ubuntu with real conviction, and still landing nowhere near what’s actually happening on the ground. That’s already starting to happen, from both directions. I’d rather name it now, plainly, while there’s still something left to protect, than write a warmer piece and watch it happen anyway.
That’s a very different conversation than “come learn from our warm community values.” And I think it’s the one worth having.
Thank you for reading!
A quick note before you go
If you’ve been following my work for a while and have been wondering how you can be more involved, there are still a few spots open for founding members of the Circle.
We’re slowly but surely growing: two months in, we have 4 Founding Members and 19 African innovators in the Circle. There’s still room for a few more Founding Members who want to support my work and its growth from the ground up, and I’d love to have you around the table from the beginning.
If that sounds like something you’d be interested in, click on the button below to fill in the form and let’s get started.
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.