The Dashboard That Hides Intersectional Harm
Identify dashboard design choices that can hide subgroup disparities during a fairness review.
The audit looks tidy. Two dashboard choices make it less honest. Which parts of the scene should Nora challenge before the review accepts the model? Overall accuracy tile Correct. A single aggregate tile can hide group-level error gaps; fairness review needs disaggregated metrics, not only a blended score. Small-group exclusion filter Correct. Filtering out small slices can erase intersectional harm. Show sample size and uncertainty instead of removing the group from the audit. Export control Not the main issue. Export can be risky if uncontrolled, but a governed export is not what hides subgroup performance in this scene. Reviewer notes…
Sign up free: one personalized lesson every day, matched to your role and goals.
Already have an account? Sign in