Every figure below traces to a government, UN, or peer-reviewed source — plus the eight insights worth saying out loud, and the claims that will get you caught.
Uganda’s binding constraint is not technology or enthusiasm — it is policy conversion. The country has been assessed four times and legislated zero times.
That gap is your strongest talking point, and unlike most interview opinions it is fully documented in official sources. Read the timeline in §2 and the figures in §3 together and the argument makes itself.
Lead with these. Each is official, citable, and checkable by whoever is interviewing you.
Ministry of ICT & National Guidance with the ITU. 2024 review, published Dec 2025. The single best document: baseline survey, gap analysis, benchmarking against Kenya, Rwanda and the US, a proposed risk framework, and seven sections of recommendations.
26 May 2026, Hotel Africana, Kampala. Uganda’s AI Readiness Assessment Report validated. Lead National Expert: Dr. Joyce Nakatumba-Nabende. The RAM process itself launched 19 Feb 2026.
The framework Uganda used. Critical distinction: UNESCO states explicitly that RAM is not a ranking exercise. More than 70 countries have implemented or are implementing it.
Oxford Insights. 195 countries, 69 indicators, 14 dimensions, six pillars. This is where Uganda’s numeric scores come from — and where the flatlining shows.
135 countries, 68,138 data points, 38 indicators, covering Nov 2023 – Sep 2025. Grounded in the UNESCO Recommendation on the Ethics of AI. Use it for the implementation-gap argument.
Adopted in Kigali, 1 April 2026. Regional AI Technologies Fund, EAC AI Alliance, a Regional Centre of Excellence for Emerging Technologies, and an explicit “African AI sovereignty” framing.
The pre-AI foundation. A 2018 national task force on the Fourth Industrial Revolution produced this strategy; AI sits inside it rather than beside it.
Kampala-based non-profit. Sunflower covers 31 Ugandan languages; the SALT corpus is 25,000 sentences across six languages, built with the Makerere AI Lab and the Makerere Institute of Languages. Open source, openly licensed.
Makerere University’s AI and data science centre, launched March 2025. Research themes: NLP for African languages, health, agriculture, climate. Home of the AI4D research lab.
NITA-U with the World Bank. National Backbone Infrastructure, the National Data Centre, e-government services, and digital inclusion covering refugees in 12 host districts and 13 settlements.
Read the right-hand column top to bottom. Assessment after assessment — and the strategy is still in draft.
Verification note: two versions of the 2025 Oxford Insights report circulate with slightly different Uganda figures (rank 113 vs 117). Confirm against the published PDF before quoting a rank aloud. The 2024 edition used 40 indicators across three pillars; the 2025 edition restructured to 69 indicators across six, so cross-year comparisons are indicative only.
Each of these is defensible from a named source. The last two are the ones that separate a prepared candidate from a well-read one.
A Policy Capacity score of 12 out of 100 sits beside Uganda’s complete absence of AI legislation. The ITU assessment states it without hedging: “An AI policy which is missing currently.” The Data Protection and Privacy Act 2019 is silent on algorithmic transparency and automated decision-making.
Source: ITU landscape assessment · Oxford InsightsGARI, UNESCO RAM, UNDP’s AIRA, and the ITU landscape assessment — and the ITU report openly says it leveraged all of them as frameworks. Uganda’s problem is not diagnosis. It is conversion.
Source: ITU landscape assessment, §5The Aeonian Project at Karuma: a 100 MW hyperscale, Tier-4+ facility drawing surplus pre-transmission power from the 600 MW Karuma hydro plant, with Nvidia and European development partners, phasing in from late 2026. Meanwhile the national AI policy sits in draft — and an estimated 98% of African data is processed abroad. Uganda is about to get sovereign compute without a sovereign AI policy.
Source: EAC / iAfrica reporting on the Aeonian ProjectSunflower covers 31 Ugandan languages; SALT is 25,000 sentences across six languages built with Makerere. This is precisely the World Bank’s 2026 World Development Report model — adopt, adapt, advance — in practice. Uganda’s genuine comparative advantage is linguistic and data sovereignty, not frontier models.
Source: Sunbird AI · World Bank WDR 2026Baguma et al. (2023), Towards an AI readiness index for Africa, argues the Government AI Readiness Index’s indicators privilege Northern research ecosystems and systematically misrepresent African institutional realities. If an interviewer cites Uganda’s ranking, you can name the critique and its authors. This is the single most impressive move available to you.
Source: Baguma et al. 2023, IDIA · cited in AI & Society, 2026GIRAI 2026 found frameworks on paper vastly outrunning implementation, especially in the Global South. Uganda’s draft-awaiting-Cabinet is a textbook instance — which lets you discuss it structurally rather than apologetically.
Source: GIRAI 2026The EAC Kigali Declaration created a Regional AI Technologies Fund, an EAC AI Alliance, and a Regional Centre of Excellence for Emerging Technologies — all framed around African AI sovereignty, with commitments to Kiswahili and regional-language models on regional infrastructure. Uganda can advance through regional instruments while national ones stall, and Uganda’s planned AI factory is named in that regional framing.
Source: EAC Declaration on AI, 1 Apr 2026Minister Baryomunsi explicitly flagged personal-data misuse ahead of Uganda’s 2026 general elections. The un-tabled AI Bill 2025 would mandate human control over AI systems; the government is instead pushing re-enactment of the Computer Misuse Act as a stopgap. Watch that substitution — it is the tell.
Source: iAfrica · Nile Post, Aug 2026Each of these is a plausible-sounding claim that a well-informed interviewer will catch.
Naming the right person for the right institution signals that you actually read the documents.
“In 1956 scientists asked whether machines could simulate intelligence; in 2026 they ask what machines should be allowed to decide.”