AI hiring monoculture is delivering racial bias at scale

Applicants who used AI to apply to multiple companies experienced higher rejection rates than expected under independent screening. Rejections for Black and Asian candidates occurred more often than baseline figures predicted. The analysis estimated that about 29,000 additional Asian candidates would have been interviewed if AI screening had not been used. Researchers raised concerns about AI screening tools because they combine widespread adoption, high impact on hiring outcomes, and public opacity. They warned that these dynamics can shape workplaces into a monoculture, which may harm companies’ future performance and adaptability.

"Applicants who applied to multiple companies using AI had all their applications rejected more often than would be expected if each company's screening methods were independent. They calculated that Black and Asian candidates were rejected in greater numbers than baseline figures would suggest. According to the survey, 29,000 more Asians would have been interviewed if AI had not been deployed."

"The researchers are concerned about the way in which AI is being used. "AI screening tools bring together three properties that should not co-exist in high-stakes decision-making: They are pervasively adopted, highly consequential, and opaque to the public," they said in a news release presenting their work."

"The effect of this will lead to workplaces dominated by a monoculture which may not be beneficial for companies going forward."

Read at Computerworld