Algorithmic Hiring Monocultures Show Racial Disparities
A study analyzing 3 million applicants and 4 million applications screened by a single algorithm vendor reveals racial disparities in hiring outcomes. Among Asian and Black applicants, 14.74% and 25.87% of applications, respectively, went to positions with adverse impact under U.S. employment discrimination standards. Additionally, 4% of applicants applying to 10 positions were uniformly rejected across all, exceeding chance expectations. The findings suggest algorithmic monoculture in hiring leads to homogeneous rejection patterns and racial bias.
Key facts
- 3 million applicants submitted 4 million applications
- All applications screened by algorithms from the same vendor
- 14.74% of Asian applicants' applications went to adversely impacting positions
- 25.87% of Black applicants' applications went to adversely impacting positions
- 4% of applicants applying to 10 positions were rejected from all
- Uniform rejection rate higher than expected by chance
- Study uses U.S. employment discrimination standards
- Algorithmic monoculture leads to homogeneous outcomes
Entities
Locations
- United States