Fairness Perceptions of Algorithmic Decision-Making: A Systematic Review of the Empirical Literature

Abstract

A comprehensive review of how people perceive fairness in algorithmic decisions across different domains. The review reveals that fairness perceptions are highly context-dependent and often diverge from technical definitions of algorithmic fairness.

Key Findings

• People use different fairness criteria depending on the decision domain
• Technical fairness metrics often conflict with lay perceptions of fairness
• Transparency about algorithmic processes does not always increase fairness perceptions

Source