Timnit Gebru

Timnit Gebru coauthored Gender Shades with Joy Buolamwini. Its intersectional evaluation examines how a headline score can conceal performance differences across groups.^[1]

Where to start

Read the paper’s task definition and labeling choices before its results. Then follow NIST’s identity-recognition evaluation and the Williams incident. Ask what each system is trying to predict, what counts as an error and how a result enters a decision.

This contribution-based entry currently follows the reviewed paper. It does not attribute all wider views, institutional roles or collaborators’ testimony to Gebru. Further original interviews and research can broaden it over time. Social Alignment connects the reading path to people affected beyond an operator.