Zum Hauptinhalt springen

Fairness Is a Choice of Harm, Not a Single Score

Explain why fairness choices depend on the harm, population, and decision context rather than on one universal metric.

data-ethics5 min read

A model can be accurate and still behave unfairly. The missing step is usually that no one said which harm they were trying to prevent. Why the metric changes with the stakes A fairness metric is a lens, not a verdict. Selection-rate parity highlights unequal access to benefits. Error-rate parity highlights uneven burdens. Calibration checks whether a score carries the same meaning across groups. If a team treats these as interchangeable, it can unintentionally protect the wrong thing. What good teams do instead Start with the decision, the people affected, and the consequence of being wrong. Name the dominant harm,…

Read the full lesson

Sign up free: one personalized lesson every day, matched to your role and goals.

← Back to library

Kontakt aufnehmen