Sourcing Bias on Job Matching Platforms: Regulation and Optimal Platform Strategy

Abstract

Recent research has found evidence of bias in algorithms used for hiring. Though recent literature on algorithmic fairness focused on improving fairness in selection algorithms, little work has focused on algorithms used in the resume sourcing phase. Algorithmic fairness in the selection phase will not solve the hiring bias problem unless bias in the sourcing phase is mitigated. This paper aims to address the bias problem in the resume sourcing phase of recruitment. We consider a job matching platform with a database of resumes belonging to two groups of jobseekers, each group varying in the proportion of relevant resumes. We identify the condition for bias in resume sourcing. We also identify the platform’s optimal strategy under different policy regulations.

Keywords: AI Bias, Job Matching platforms, Regulation

JEL Classification: M15