AI RESEARCH

Decomposing Discrimination: Causal Mediation Analysis for AI-Driven Credit Decisions

arXiv CS.LG

ArXi:2603.27510v1 Announce Type: new Statistical fairness metrics in AI-driven credit decisions conflate two causally distinct mechanisms: discrimination operating directly from a protected attribute to a credit outcome, and structural inequality propagating through legitimate financial features. We formalise this distinction using Pearl's framework of natural direct and indirect effects applied to the credit decision setting.