Algorithmic Bias in Hiring

How bias gets in

If a hiring model learns from past decisions, it can learn past patterns too, including unfair ones. A well-known example: reporters found in 2018 that an experimental resume-screening tool had learned to downrank resumes mentioning women's organizations, and the project was scrapped.

Proxies

Even without asking about race or gender, a system can rely on stand-ins (proxies) like zip codes, names, school names, or employment gaps. Those stand-ins can link to protected characteristics.

What the law says

Federal anti-discrimination laws such as Title VII, the ADA, and the ADEA apply no matter whether a decision is made by a person or software. The ADA also gives you the right to ask for reasonable accommodations on assessments.

The EEOC used to publish technical-assistance documents explaining how these laws apply to AI, but it removed them from its website in January 2025. That guidance never created the obligations; the laws themselves still apply. Some states and cities have their own AI-specific rules. For example, New York City's Local Law 144 requires bias audits for certain automated employment decision tools.

What you can do

You can ask an employer whether AI is used in its process and how to request an accommodation. Keep records of your applications. If you believe you were treated unfairly, the Fair Hiring Center lists official channels.