This is part 1 of a five-part series on artificial intelligence — reading historian Yuval Noah Harari's recent concepts through Rwanda's own experience of building institutions from scratch. ALSO READ: Kagame says investing in Africa's AI solutions a win for everyone Emmanuel has been riding his taxi-moto in Kigali for nine years. Last month, a passenger pulled up her route on her phone and told him to follow it. He looked at the screen, shook his head, and took a different road anyway. Roadworks near Kimisagara, he said. The App didn't know that yet. ALSO READ: What to know about new global commission on AI co-chaired by Kagame That small refusal is the whole story of artificial intelligence, if you look at it the right way. We've spent years telling ourselves AI is just a tool — a faster calculator, a smarter radio, a coffee machine with better manners. A tool waits. It does exactly what you tell it and nothing more. Your washing machine does not decide what you should wear tomorrow. ALSO READ: How Rwanda is regulating artificial intelligence The historian Yuval Noah Harari has a name for what's changed: AI isn't a tool anymore, it's an agent. It makes calls nobody programmed it to make, and drifts toward conclusions even its own creators can't always retrace. Not conscious, necessarily. Just deciding. The machine that learned to surprise Go back to May 1997, a hotel conference room in Manhattan. Garry Kasparov, the best chess player alive, sat across from a machine called Deep Blue. IBM had fed it a century of grandmaster wisdom — every opening, every known trap. Kasparov expected a strong opponent. In the second game, Deep Blue made a quiet positional move no textbook recommended and no engineer had planned for. Kasparov was so unsettled he assumed a human must be helping it. He lost the match days later, and insisted for years afterward that something had happened in that room nobody could fully explain. What people missed is that Deep Blue hadn't just memorized chess — it had reasoned past its own training, arriving at moves that surprised the people who built it. That was the real headline in 1997, and almost nobody wrote it that way. The story became computer beats man, a simple contest with a winner and a loser. The deeper story, that a machine had crossed from executing instructions to generating its own judgment, got buried under the scoreboard. Twenty-nine years later, still buried The same mistake is happening again, quietly, in ordinary places. Dispatch apps, chatbots, farming advisories, loan approvals — most people still talk about them the way 1997 talked about Deep Blue before the tournament: useful, impressive, ultimately just following orders. We keep mistaking a mind for a mechanism because the mechanism used to be all there was. Here's the uncomfortable part. Emmanuel ignoring the app wasn't him beating the machine. It was him doing the one thing it still can't: noticing something new about the world and deciding what to do about it, without anyone telling him how. That gap is shrinking fast. The systems being built right now aren't trying to out-navigate a taxi-moto rider in Kimisagara. They're trying to out-decide entire ministries, banks, hospitals — the loan officer who used to override a bad score, the triage nurse who used to trust her own eyes — institutions that, unlike Emmanuel, rarely notice when their judgment has quietly been replaced. Rwanda has spent three decades rebuilding institutions from the ground up, which means Rwandans understand something most countries have forgotten: institutions are only as good as the judgment inside them. A country that doesn't ask who, or what, is making the actual decisions is a country that has already handed them over. The question isn't whether AI will become more capable. It already has. The question is whether we'll notice the moment it stops waiting for instructions and starts writing its own. Emmanuel noticed the roadworks before his app did. The next version might not give him the chance. Next week, in part 2, we look at the natural habitat of this new technology — why AI thrives on modern bureaucracy and spreadsheets the way fire, once just a danger, became the thing every civilization was eventually built around, and what that means for a nation built on institutional systems. The writer is a professional engineer specializing in process design.
