Dr. Muganga,
Thank you for your letter explaining your vision for AI to President Museveni. It deserves a serious response — not nearly the dismissive social media reactions it has garnered, but a practitioner's full engagement with the arguments you are making. I intend to give you that.
I have been part of Uganda's technology story for over two decades — not as a donor or an observer, but as a builder. I am one of the founding members of the ICT Association of Uganda. I worked inside the Ministry of Health digital health working groups during the era when UNICEF's technologists coined the term "pilotitis" to describe exactly what was happening to Uganda's eHealth ecosystem — a term born from our specific national experience of good intentions colliding with broken implementation architecture. I deployed community health information systems at national scale across six countries. I founded Guild Digital, which gave me every hard lesson I now draw on.
I am also not here to challenge your credentials. I am doing something more specific: running a project feasibility check. My job — literally, every day — is to ask whether the conditions for success exist before resources are committed. That discipline requires having been in enough rooms where confident projections met unforgiving implementation reality.
Where We Agree
The AI-driven wealth creation opportunity is real. Uganda's demographic profile — more than three-quarters of its population under 34 — is a genuine strategic asset, not a liability. The urgency you feel is legitimate. The nations that build AI capability in this decade will not simply grow faster; they will reshape the terms on which the rest of the world trades, borrows, and governs. Sitting out is not a neutral position. On all of that, I stand with you.
Where I part ways is on the mechanism. In technology implementation, the mechanism is everything. A wrong mechanism does not merely delay the outcome. It consumes the capital and political will needed to build the right one.
On the Economics
Uganda's nominal GDP stood at $56.6 billion in 2024, with a GDP per capita of $1,208. The country has averaged real GDP growth of 5.1 percent over the last decade. That is a creditable record. It is not the foundation for sustaining 14 to 18 percent annual growth for 14 consecutive years, which is what your $500 billion target requires. Not one sub-Saharan African economy has done this at scale. The arithmetic is not just ambitious — it is unmoored.
The projections you cite come primarily from ARK Invest — a firm whose entire business model is built on best-case technological adoption curves. ARK's flagship calls on Tesla, Bitcoin, and genomics have consistently exceeded actual outcomes by significant margins. Citing ARK to the President is not analysis. It is advocacy dressed in the language of data, and Uganda's transformation deserves better scaffolding than that.
In technology implementation, the mechanism is everything. A wrong mechanism does not merely delay the outcome. It consumes the capital and political will needed to build the right one.
On the One-Laptop-Per-Child Proposal
The one-laptop-per-child idea has a documented international track record, and it is not encouraging. The OLPC programme, which distributed millions of laptops across developing countries from 2005 onward, produced learning outcomes that were, by the programme's own assessments, largely disappointing. The devices arrived in schools without reliable electricity. They arrived without trained teachers. They arrived without internet connectivity. They arrived without maintenance contracts. They became expensive paperweights at a rate that would embarrass any serious implementation review.
Uganda's Net Enrollment Rate in primary education is approximately 83 percent. Completion rates are significantly lower. The constraint is not hardware. The constraint is teaching quality, infrastructure, and the economic pressure on families that pulls children out of school before they complete. A $1.2 billion laptop distribution programme does not address any of these constraints. It adds a device to a system that lacks the conditions to use it.
The Pilotitis Problem
I want to be specific about the implementation architecture risk, because I have watched it destroy well-funded programmes firsthand. Uganda's eHealth ecosystem, as of 2015, had over 40 parallel digital health initiatives running simultaneously — each with its own data standard, each with its own donor, each with its own implementation team, and virtually none of them interoperable. The UNICEF technologists who coined "pilotitis" were describing this exact phenomenon: the compulsive launching of technology pilots that never scale, consume enormous resources, and leave a trail of orphaned systems behind them.
A $1.2 billion national AI programme, announced before the institutional capacity to absorb it exists, will not avoid this pattern. It will repeat it at scale. The history of ICT4D is littered with high-profile, well-intentioned programmes that failed not because the technology was wrong, but because the implementation architecture was missing.
A Credible Alternative
Here is what I would propose instead — lower cost, higher fidelity, and executable within Uganda's actual institutional capacity:
Phase 1 ($50–80M over 3 years): Build AI infrastructure at five anchor institutions — Makerere University, Uganda Management Institute, and three sector-specific technical colleges. Not laptops distributed broadly, but high-specification computing environments where genuine AI development can happen. Fund 200 AI research fellowships for Ugandan nationals, with a mandatory return-service requirement. Establish data governance legislation.
Phase 2 ($100–150M over 5 years): Deploy AI applications in three sectors where Uganda has both genuine data assets and acute need: agriculture (crop disease detection, yield prediction), health (disease surveillance, diagnostic support), and financial services (credit scoring for the unbanked). These are not speculative applications. They are deployable today with existing technology and existing Ugandan data.
Phase 3 (revenue-generating from Year 4): License the infrastructure, the trained talent, and the proven applications to regional governments and international development organisations. East Africa has a $2.3 billion development finance market. Uganda can capture a meaningful share of it as a credentialed AI implementation partner — but only if the first two phases are executed with discipline.
Total cost: approximately $200–250 million over eight years. Roughly one-fifth of your proposed budget. Implementation complexity: substantially lower. Probability of producing measurable outcomes: substantially higher.
The Honest Question
The honest question your proposal raises is not whether Uganda should invest in AI. It should. The honest question is whether a $1.2 billion programme announced before the institutional preconditions exist is a development strategy or a political signal. I am not in a position to judge your intentions. I am in a position to assess the mechanism. And the mechanism, as proposed, will not produce the outcomes you have promised.
Uganda's young people deserve the AI future you are describing. They deserve it built on a foundation that can actually hold. A bad road, however confidently travelled, cannot lead us home.
With respect,
Brian Ssennoga