Analysing documented failure cases

Link real, published failures to the principle that failed and the metric that would have caught it.

BeginnerResponsible AI Lab
العربية

The idea

Four cases with sources; choose the principle and compare with the analysis.

Data description

Summaries of cited published sources.

Suggested next experiment

Scientific sources for this experiment

  1. Machine BiasJ. Angwin, J. Larson, S. Mattu, L. Kirchner (ProPublica) (2016) — Documented case
  2. Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification (PMLR 81:77–91)J. Buolamwini, T. Gebru (2018) — Paper
  3. Simple Demographics Often Identify People UniquelyL. Sweeney (2000) — Paper

More in this module