Study Finds Racial Disparities and Homogeneous Rejection Patterns in Algorithmic Hiring Systems

Summary

This research paper investigates the phenomenon of "algorithmic monoculture" in hiring, where many employers use screening algorithms from the same few vendors. The authors analyze a novel dataset of 3 million applicants and 4 million applications screened by a single vendor's algorithms. They find significant racial disparities: 14.74% of Asian applicants' and 25.87% of Black applicants' submissions go to positions that adversely impact those groups under U.S. employment discrimination standards. Additionally, 4% of applicants who apply to 10 positions are recommended for rejection from all positions—higher than expected by chance. The study uses the deterministic replicability of hiring algorithms to show that applicants would need to apply widely to ensure human consideration.

Source

bskyStudy Finds Racial Disparities and Homogeneous Rejection Patterns in Algorithmic Hiring Systemsarxiv.org