Societal biases reinforcement through machine learning: A credit scoring perspective
share
Does machine learning and AI ensure that social biases thrive ? This paper aims to analyse this issue. Indeed, as algorithm are informed by data, if these are corrupted, from a social bias perspective, good machine learning algorithms would learn from the data provided and reverberate the patterns learnt on the predictions related to either the classification or the regression intended. Therefore, if the data sets are capturing the way society behaves whether it is positive or negative, then this would be reflected by the models. Using credit scoring data sets as provided by financial institutions in the US, our objective is to assess how much social biases are transmitted from the data into the scoring approach. Relying on Random Forests and SVM approaches, we details the results obtained and measure how much these biases can impact people's access to loans.
READ FULL TEXT