Domain: Human Resources/Employment
Description: Amazon developed an AI recruitment tool (2014-2017) that systematically discriminated against women, leading to its termination. Recent University of Washington studies (2024) found similar bias in LLMs for resume screening.

Ethical Challenges:

• Gender Bias: Amazon's system penalized resumes containing words like "women's" and downgraded graduates from women's colleges
• Intersectional Bias: UW study found names associated with Black women were never favored over white men
• Historical Inequality Perpetuation: AI trained on historical workforce data predominantly male
• Disparate Impact Discrimination: Under Title VII violations

Public Datasets:

• Primary: Adult/Census Income Dataset - UCI Machine Learning Repository
• URL: https://archive.ics.uci.edu/dataset/20/census+income
• MIT Case Study: https://ocw.mit.edu/courses/res-ec-001-exploring-fairness-in-machine-learning-for-international-development-spring-2020/pages/module-four-case-studies/case-study-mitigating-gender-bias/
• Content: 48,842 records of census data for predicting income >$50K/year with 13 features
• Bias Use: Widely used for studying bias in income prediction across demographic groups
• Additional Platforms: Available on fairness toolkits - https://fairmlbook.org/datasets.html