Protecting the Protected Group: Circumventing Harmful Fairness

Machine Learning (ML) algorithms shape our lives. Banks use them to determine if we are good borrowers; IT companies delegate them recruitment decisions; police apply ML for crime-prediction, and judges base their verdicts on ML. However, real-world examples show that such automated decisions tend to discriminate against protected groups. This potential discrimination generated a huge hype both in media and in the research community. Quite a few formal notions of fairness were proposed, which take a form of constraints a “fair” algorithm must satisfy. We focus on scenarios where fairness is imposed on a self-interested party (e.g., a bank that maximizes its revenue). We find that the disadvantaged protected group can be worse off after imposing a fairness constraint. We introduce a family of $\mathit{\text{Welfare-Equalizing}}$ fairness constraints that equalize per-capita welfare of protected groups, and include $\mathit{\text{Demographic}}\mathit{\text{Parity}}$ and $\mathit{\text{Equal}}\mathit{\text{Opportunity}}$ as particular cases. In this family, we characterize conditions under which the fairness constraint helps the disadvantaged group. We also characterize the structure of the optimal $\mathit{\text{Welfare-Equalizing}}$ classifier for the self-interested party, and provide an algorithm to compute it. Overall, our $\mathit{\text{Welfare-Equalizing}}$ fairness approach provides a unified framework for discussing fairness in classification in the presence of a self-interested party.