Fairness under Competition

TL;DR

This paper examines how individual fairness in classifiers can lead to unfair outcomes at the ecosystem level when firms compete, highlighting the potential negative effects of fairness adjustments on overall system fairness.

Contribution

It introduces a novel study of fairness in competitive ecosystems, showing that individually fair classifiers can result in unfair collective outcomes, supported by theoretical and experimental analysis.

Findings

Fair classifiers can lead to unfair ecosystems under competition.

Adjusting classifiers for fairness may decrease overall ecosystem fairness.

The impact depends on classifiers' correlation and data overlap.

Abstract

Algorithmic fairness has emerged as a central issue in ML, and it has become standard practice to adjust ML algorithms so that they will satisfy fairness requirements such as Equal Opportunity. In this paper we consider the effects of adopting such fair classifiers on the overall level of ecosystem fairness. Specifically, we introduce the study of fairness with competing firms, and demonstrate the failure of fair classifiers in yielding fair ecosystems. Our results quantify the loss of fairness in systems, under a variety of conditions, based on classifiers' correlation and the level of their data overlap. We show that even if competing classifiers are individually fair, the ecosystem's outcome may be unfair; and that adjusting biased algorithms to improve their individual fairness may lead to an overall decline in ecosystem fairness. In addition to these theoretical results, we also provide supporting experimental evidence. Together, our model and results provide a novel and essential call for action.