A notable case of bias in AI recruitment involved Amazon's automated hiring tool, which was
found to discriminate against women. This tool was developed to mechanize the hiring
process by reviewing resumes and identifying top candidates. It trained on data from the
company's resumes over a 10-year period, which mostly came from men due to the male
dominance in the tech industry. As a result, the AI learned to prefer male candidates,
penalizing resumes that included words like "women's" (e.g., "women's chess club captain")
and downgrading graduates from all-women's colleges. Despite efforts to neutralize these
biases, the tool continued to show discriminatory tendencies, leading Amazon to eventually
discontinue its use​​​​. LINK
One illustrative case study involved a financial institution, Money Bank, which utilized an AI
tool called GetBestTalent for shortlisting candidates. Despite assurances of bias and
discrimination audits, several candidates—distinguished by gender, race, and
age—suspected that their rejections were unlawfully discriminatory. These suspicions arose
because AI's decision-making process, which often relies on opaque algorithms, made it
difficult to understand why certain candidates were rejected​​. LINK
In another notable example, the company HireVue developed software that assessed job
applicants based on their facial movements, word choice, and how they spoke. This instance
highlights how AI tools can embed biases based on the subjective criteria they're
programmed to evaluate, further complicating the ethical landscape of AI in recruitment​​.
LINK