Back to IEEE CertifAIEd
IEEE CertifAIEd · Algorithmic Bias
Algorithmic bias
“Identification, measurement, and mitigation of unjustified disparate outcomes.”
CertifAIEd Algorithmic Bias criteria
datamodeldeploymentmonitoring
Type
criterion
Parent
—
Sub-items
2
Sub-items
Cross-framework relationships
Lattice uses an explicit, four-type vocabulary and refuses ‘equivalent.’ Each relationship below carries a rationale and an editorial reviewer credit.
relates to
1- MeasureNIST AI RMF · MEASURE
NIST MEASURE practices include bias measurement aligned with CertifAIEd's bias criterion; CertifAIEd adds certification-grade evidence requirements.
Reviewer: Editorial: A. Vega · Reviewed 2026-04-22
partially addresses
1- Fairness and non-discriminationUNESCO · Fairness & non-discrimination
CertifAIEd's algorithmic bias criterion partially addresses UNESCO fairness; UNESCO covers broader social justice considerations beyond model-level disparity.
Reviewer: Editorial: A. Vega · Reviewed 2026-04-22
Version: Program v2Last reviewed: 2026-04-12