AI Hiring Tools Consistently Amplify Bias Against Women and Minorities, Studies Show

Automated CV screening systems are increasingly making hiring decisions before any human recruiter reviews an application, but mounting evidence suggests these tools reinforce rather than reduce discrimination. Machine-learning models trained on a company's historical hiring data learn to replicate past patterns, effectively encoding existing workforce skews by gender, race, age, or class into their scoring. Amazon's widely cited case from 2018 showed its experimental CV-screening tool penalised résumés mentioning women's organisations and all-women's colleges, leading the company to scrap the project entirely. A 2024 University of Washington study by researchers Kyra Wilson and Aylin Caliskan tested leading language models across over three million CV comparisons and found a consistent preference for candidates with white-associated names. Experts argue that algorithmic bias in hiring is not an isolated technical flaw but a structural outcome of how these systems are designed and trained.