Group Fairness

Fairness criteria applied to groups defined by protected attributes, ensuring that statistical measures of outcomes are equitable across groups.

Looking at fairness for entire groups rather than individuals; making sure outcomes like approval rates, error rates, etc. are roughly equal across groups defined by race, gender, age, etc.

Group fairness metrics are foundational for AI compliance reporting. Your governance framework should track these metrics across all high-risk AI systems and establish clear escalation procedures when thresholds are breached.