Research reveals stark racial bias in AI resume screening systems.
A University of Washington study found that large language models screening resumes favored white-associated names 85% of the time, and never once favored a Black male name over a white male one. This bias extends beyond names on paper to voices in the room. Speech models like Whisper produce significantly higher error rates for non-native and accented speakers because the audio they trained on skews American and native. The AI inherited our blind spots because it learned from us.
— If Your Name Isn’t Western, AI Could Cost You The Promotion. · Forbes