Research Focus
The Gender Data Gap in African AI Systems
Examining how underrepresentation in training data perpetuates inequality in algorithmic outputs across the continent.
Why it matters
Across Africa, women remain systematically underrepresented in the datasets that train modern AI systems — from facial recognition and credit scoring to health diagnostics and labor-market platforms. When women are missing from the data, they are effectively missing from the decisions those systems shape. The result is a quiet but compounding form of digital exclusion that no amount of post-hoc fairness tuning can fully correct.
What we are studying
Our research traces the full pipeline — from data collection practices in African institutions, to labeling and annotation workflows, to deployment of AI systems in finance, health, and public services. We document where gender gaps emerge, quantify their downstream effects on women's access to opportunity, and test interventions that close them at the source rather than at the model.
What we are building toward
Evidence that African data stewards, regulators, and AI builders can use to design gender-equitable systems from the ground up: dataset audits, procurement standards, and capacity programs that put women's data — and women data scientists — at the center of the continent's AI future.
Interested in this work?
Reach out to collaborate, fund, or learn more.