Beneath the vibrancy of our digital innovation narrative lies a deeper, more uncomfortable truth, one that we have been reluctant to confront with the seriousness it demands. As artificial intelligence becomes the new foundation upon which economies, governance systems and social interactions are built, Kenya’s position within this global order reveals a structural vulnerability, a quiet erosion of sovereignty that resembles a new form of colonization. Not one enforced through territorial conquest or political domination, but through the infrastructures, datasets, cloud platforms and algorithmic systems that increasingly mediate every aspect of our national life.
We do not have control over the core layers of the AI ecosystem that it depends on. Our universities, startups, government agencies and research institutions rely on cloud servers hosted abroad, GPUs manufactured in distant economies, proprietary models trained on data we do not own and platforms that impose policies without negotiation. Our digital public services run on infrastructures we rent. Our innovation ecosystem depends on APIs governed by terms of service that can shift overnight. Even our cultural and linguistic knowledge, the rich content produced daily by millions of Kenyans, flows into global datasets that train frontier models far outside our borders, only for the insights to be resold to us as premium AI products. It is a subtle but powerful form of extractivism that is data leaves Kenya as a raw material and returns as a finished product owned by someone else and we buy it.
Much of what appears to be “Global South innovation” is built upon stacks that we neither governs nor fully understands. A typical solution praised as local may have a local interface and local branding, but beneath the surface it rests on foreign platforms, foreign algorithms, foreign storage systems and foreign payment rails. What we call digital entrepreneurship often amounts to integration work performed inside infrastructures designed elsewhere. The more deeply we digitize, the more tightly we bind ourselves to the architectures of others. It becomes difficult to speak meaningfully of autonomy when the intelligence layer, the decision-making layer, the computational layer and the data layer all sit outside our control.
This phenomenon is not abstract. It has very real implications for how we govern ourself. When AI systems begin influencing public policy, credit scoring, recruitment, policing (even in the security sector), education and health, imported algorithmic logic becomes an invisible layer of governance. Systems that do not understand local cultures, languages, socio-economic realities or moral frameworks begin shaping outcomes for millions of people. The danger is not only external but also it is internal as well. Even if we were to build full technological sovereignty, without accountability and democratic oversight we could still reproduce colonial power structures within our own borders, where a technologically empowered elite governs through opaque digital systems while the rest of society becomes data subjects rather than citizens.
The illusion of control in Kenya comes from our visible activities such as hackathons, innovation hubs, AI labs, mobile apps, digital payments, smart agriculture pilots, digital government platforms. At first glance, these paint the picture of a country at the forefront of African innovation. Yet innovation is not defined by how many apps we build, but by how deeply we control the infrastructures that make intelligence possible. A country that trains students merely to operate imported models and assemble foreign platforms cannot meaningfully claim sovereignty in the age of AI. We may celebrate M-PESA as a local miracle, yet much of the underlying security architecture, analytics tooling and cloud infrastructure rests on external systems. We may celebrate digitization in government, yet critical datasets remain hosted abroad or processed through external tools. We may celebrate the growth of Kenyan EdTech, AgTech and FinTech, yet the intelligence behind these innovations is often powered by models built in California.
When viewed through the lens of history, the parallels become difficult to ignore. Classical colonialism extracted physical resources such as land, minerals, labor. Today’s digital colonialism extracts data, attention, cultural patterns and linguistic structures. The former justified itself through the rhetoric of “civilization”. The latter disguises itself through the language of “innovation”, “efficiency” and “digital transformation”. The intentions may differ, but the structural consequences are similar: dependency, asymmetry and loss of agency.
Ultimately, the most critical layer of sovereignty is cognitive. A nation’s control over its technological destiny begins with how its people are educated to think about technology. If we train a generation of developers who merely assemble external tools, we will remain digitally dependent. But if we train creators, theorists, policymakers, linguists, ethicists and systems thinkers and individuals who can interrogate, question, redesign and imagine alternatives and then Kenya will build its own center of gravity within the AI world.
The question “With AI, do we even have control over anything?” is therefore not a lament but a provocation. It forces us to confront the uncomfortable reality of our digital dependencies and to design pathways out of them. It asks us to imagine a country where our data is governed by us, our languages are modeled by us, our infrastructures reflect our values and our destinies are shaped not by the silent rules of external algorithms but by our own collective intelligence.
