Individualisation, scoring and fairness : some thoughts based on credit and insurance

Department of Law & Tax
Speaker : Nathan GENICOT (ULB)
Room S126

Abstract: “Recent progress in computing capacity and machine learning, as well as new forms of data collection, have given rise to new practices of quantifying individual behaviour, notably in the form of scoring systems. These are regularly denounced as unfair, particularly when they are based on protected criteria likely to lead to discrimination. In particular, they are criticised for not assessing subjects individually, but solely on the basis of their membership of certain groups. However, this criticism is likely to miss the point, as scoring systems are intended to make a judgement that is as individualised as possible (and do so all the better as they are fed by a lot of data). This presentation will explore this tension by looking at some cases in the credit and insurance fields.”