Building evidence and accountability around facial-analysis systems
From the archive. Deadlines and opportunity details reflect what was reported at the time.
From the builder archive: a previously researched story, with achievements dated in the profile.
Country / heritage: Ghana
Joy Buolamwini’s builder story combines technical research with an organization focused on accountability in artificial intelligence. Born in Canada to parents from Ghana, she developed a research and advocacy career examining how automated systems perform across different groups of people. In 2018, Buolamwini and Timnit Gebru published Gender Shades, evaluating three commercial systems that classified gender from facial images. MIT’s coverage reported substantial differences in error rates across skin-tone and gender groups in the systems tested at that time. The research made a broad concern measurable through a defined audit and published evidence. Buolamwini also founded the Algorithmic Justice League, bringing together research, creative work and public education around AI harms. Africa Giant featured her work in October 2026. For African technology builders, the practical relevance lies in testing a product against the people who will actually use it. The 2018 findings describe the tested historical systems and are not presented as measurements of current software.
Builder lesson: Evaluate AI products with representative users and publish clearly bounded findings; an overall accuracy figure can conceal important performance differences.
Sources: Source 1 · Source 2 · Source 3
Profile research checked: 2026-10-26