
Episode #66
Africa's 800+ languages and the AI data gap
Dr. John Mugo is the Chief Executive Officer of Mizizi Elimu Afrika, the African-led education organization based in Nairobi, Kenya, formerly known as Zizi Afrique Foundation. He holds a Doctorate and a Master's degree in Special Needs Education from the University of Hanover, Germany, and a Bachelor of Education (Music) from Kenyatta University, where he went on to chair the Department of Special Needs Education. Before founding the organization in 2018, he headed Uwezo, one of East Africa's largest citizen-led learning assessment initiatives, measuring what children actually know rather than what systems claim to teach. He leads ALiVE (Action for Life Skills and Values in East Africa), a regional framework and assessment system that has now assessed life skills and values in over 45,000 adolescents across Kenya, Uganda and Tanzania, and works directly with the Kenya Institute of Curriculum Development on values-based education and teacher training reform. In March 2026, Zizi Afrique rebranded as Mizizi Elimu Afrika β "mizizi" is Kiswahili for "many roots" β and launched Vision 2040: Strong Foundations for Lasting Change, a 15-year strategy to reach more than 10 million children across 10+ African countries with stronger foundational learning. His organization currently works across Kenya, Tanzania, Uganda, Rwanda, Malawi and Cameroon, and is piloting AI-supported teaching with Fab AI and Gemini in 15 Kenyan schools. In this episode, John talks with Svenia Busson about: Why fewer than 20% of children in much of Sub-Saharan Africa are learning at grade level β and why he calls that an "80 to 90% leakage" no amount of university investment can fix Vision 2040 and the two-lever theory behind it: strengthening government systems instead of building parallel NGO programmes, and forcing a competitive civil society sector to collaborate Why Africa's real AI problem isn't models but data β 800+ languages, oral cultures, and knowledge trapped in unpublished grey literature The case for voice data and mother-tongue AI as the fastest route to reading β drawn from his own experience learning to read in Kikuyu, not English What the Fab AI and Gemini pilot in 15 schools is showing about real-time assessment β and the widening gap between connected private schools and public schools without electricity The whole-school approach to values built with Kenya's curriculum institute and piloted in 79 schools β because responsibility and integrity can't be taught in a lesson slot Why place value in Grade 1 explains Kenya's KCSE maths results, and the 10-year roadmap to move Grade 3 numeracy proficiency from 30% to 75% His blunt advice to ministries drafting AI policy: stop banning it, start investing in your own data, and put young people in the driver's seat Why Africa's median age of 19 is an asset rather than a fiscal crisis β and what has to be true for that to hold Links: Mizizi Elimu Afrika Β· ALiVE Β· Uwezo Β· Fab AI Β· AI for Education Β· KICD Β· John Mugo on LinkedIn Mentioned in this episode: For those who wish to know more about the Education Futures fellowship: https://educationfutures.ai/fellowship






