Judgment Bias
Systematic errors in human or AI decision‐making processes caused by cognitive shortcuts or flawed data, requiring bias audits and mitigation.
Definition
Errors stemming from cognitive heuristics (availability, anchoring) in human judgments or from algorithmic patterns mirroring those biases. Judgment bias can occur in label annotation, policy-setting, or post-model review. Governance addresses it through structured decision protocols, blind-review panels, bias-awareness training, and algorithmic audits to detect and correct skewed outcomes before they propagate.
Real-World Example
In loan-underwriting, underwriters tend to favor applicants with familiar-sounding names (recognition bias). The bank introduces blind applications (masking names) and conducts bias audits comparing approval rates before and after implementation—reducing name-based disparities by 60%.
.svg)
