3. Algorithmic fairness

Algorithmic fairness is a term describing the conceptualization, and remediation, of bias in rule-based procedures for making decisions about different types of individuals. At heart is a presumption that some ways of discriminating between people are undesirable (i.e., “unfair”), while others are acceptable or even desirable (i.e., “permissible”). This field has quickly emerged as an active, important, and multidisciplinary research agenda driven by the increased use, and sophistication, of algorithms that affect people on an everyday basis. In this chapter, we discuss the emerging literature on algorithmic fairness and some of its connections to public choice, political economy, and social choice.