The increasing reliance on algorithmic hiring is raising concerns about diversity and inclusion in the workforce. Over 90% of U.S. employers now use algorithms to screen job applicants, with many relying on the same few vendors. This has led to what researchers call an algorithmic monoculture, where homogeneous outcomes are prevalent, and diversity is at risk.
The Rise of Algorithmic Monocultures
A recent study involving 3.4 million job applications across 156 employers highlights the extent of this issue. The study found that over 60% of Fortune 100 companies use HireVue's algorithms. This widespread use of a single vendor's technology creates a bottleneck in job opportunities, particularly affecting racial diversity.
Impact on Racial Diversity
The study reveals significant racial disparities in hiring outcomes. Black applicants are disproportionately affected, with 30% applying to positions that demonstrate adverse impacts against them. Asian applicants face the largest shortfall, with 29,000 additional applications needed to match the selection rates of the most favored racial group.
"Algorithmic monoculture could lead to systemic rejections, where applicants are rejected everywhere," the study warns.
Systemic Rejection Rates
Applicants submitting multiple applications often face systemic rejections. The study found that 10% of those submitting four applications are rejected across the board, a rate significantly higher than expected under independent decision-making.
| Racial Group | Adverse Impact | Additional Applications Needed |
|---|---|---|
| Black | 30% | N/A |
| Asian | Largest shortfall | 29,000 |
Regulatory and Legal Context
The use of hiring AI is governed by laws such as Title VII of the US Civil Rights Act, which addresses discrimination in hiring. Despite existing regulations, the study suggests that current oversight may not fully address the nuances of algorithmic hiring.
Moving Forward
To mitigate these issues, companies need to diversify their algorithmic vendors and ensure that their hiring practices are inclusive. This may involve adopting more transparent algorithms and conducting regular audits to identify and rectify biases.
