Machine Learning Credit Scores And Disparate Impact Theory

This Note analyzes the effects of machine learning in the lending context and argues that the existing legal framework can address unintentional discrimination that may result from credit-scoring models developed through machine learning. Potential liability stems from increased complexity of machine-learning processes; as machine-learning algorithms become more sophisticated, it becomes more difficult to explain the results they produce. Under current law, the inability to reasonably explain or even discover the correlations between data inputs and the resulting disparate impact leaves the lender vulnerable to suit for unintentional discrimination.