Timnit Gebru

Timnit Gebru is a computer scientist whose research examines both machine-learning systems and the conditions under which they are built. Born in Ethiopia, she later studied in the United States and earned her doctorate at Stanford University. Her technical background includes computer vision, which became a foundation for her work on the social consequences of AI.

With Joy Buolamwini, Gebru co-authored Gender Shades, a 2018 study of commercial gender-classification systems. The study measured differences in error rates across groups defined by skin type and gender, showing why aggregate accuracy can obscure unequal performance. It contributed concrete evidence to discussions of fairness in computer vision.

Gebru also co-authored Datasheets for Datasets, proposing systematic documentation of how datasets are assembled, what they contain, and how they should be used. She worked at Microsoft Research and later co-led Google’s ethical AI team. Her research connected technical evaluation with organizational practices that influence what evidence is collected and how risks are recognized.

In 2021, she founded the Distributed AI Research Institute, known as DAIR. Its work emphasizes independent, community-informed research on AI and its effects. Gebru also co-founded Black in AI, supporting the participation of Black researchers. Her contribution combines empirical studies, documentation practices, and institution building around whose experiences count in the design and evaluation of AI.