The paper starts from a situation of information asymmetry on credit markets and zooms in on AI-enhanced credit scoring as an institutional response. It assumes the potential for expanding access to credit as well as the risk of discriminatory treatment of historically disadvantaged communities. Against this background, the paper explores legal requirements of „explainability“, using two recent European Court of Justice decisions as illustrations. The paper gives an overview of XAI methods along with their socio-technical and legal limits. It contributes to the discussion by suggesting to treat explanations as a public good and designing an intermediary institution which would act as a go-between connecting consumer data subjects and scoring companies.
White Paper No. 116
