🌍
AI in Africa
7 articles
A 125M-Parameter Model for 11 Languages Is Not a Small Story
UCT researchers built an open corpus and a 125M-parameter decoder-only model covering all 11 official South African languages, nine of them low-resource. The corpus is the real contribution, and the size tells you why.
Ibrahim Denis Fofanah·Jul 31, 2026·6 min
The Broken Ruler: Why "LLMs Now Translate African Languages" Doesn't Mean What You Think
A new study on Hausa and Fongbe translation found that for Hausa, human evaluators preferred a system that no automatic metric ranked first — and BERTScore collapsed entirely. Before you trust a rising leaderboard score in any African language, check the ruler.
Ibrahim Denis Fofanah·Jul 28, 2026·7 min
Rwanda Is Betting It Can Lead African AI Without Owning the Compute
Rwanda is first in Africa on AI policy, agency and summits, but the GPUs are going to South Africa, Egypt, Kenya, Morocco and Nigeria. Inside a small country's bet on governing AI instead of hosting it.
Ibrahim Denis Fofanah·Jul 28, 2026·7 min
Sierra Leone Is Writing Its First AI Strategy. The Minister Asked a Better Question Than Most Rich Countries Do.
MoCTI and the World Bank are building Sierra Leone's first National AI Strategy on a Compute, Capacity, Context diagnostic. The framing is genuinely good. The binding constraint is that 7 million Sierra Leoneans are offline.
Ibrahim Denis Fofanah·Jul 21, 2026·7 min
AI in Africa in 2026: The Real State of Play
A practitioner's map of AI in Africa in 2026: where the money flows, who is building, and why compute and data sovereignty decide the rest.
Ibrahim Denis Fofanah·Jul 18, 2026·6 min
Google's AI Tutor Raised Maths Scores in Sierra Leone. It Raised Them Most for Students Who Were Already Ahead.
Google ran a real randomised trial of an AI tutor in Port Loko: 1,763 students, a proper control group, independently marked. Maths scores rose. But buried in Google's own report is a finding nobody is quoting, it helped the strongest students most.
Ibrahim Denis Fofanah·Jul 13, 2026·12 min
I Built the First ML Model for Sierra Leone's Rice Harvest, and It Failed
No model beat naive persistence on crop statistics alone. Free satellite data cut forecast error by a third, and May–June rainfall predicts the harvest months ahead.
Ibrahim Denis Fofanah·Jul 13, 2026·7 min