Algorithmic Fairness in Recruitment: Designing AI-Powered Hiring Tools to Identify and Reduce Biases in Candidate Selection
Citation. Agbasiere, C. L., & Nze-Igwe, G. R. (2025). Algorithmic Fairness in Recruitment: Designing AI-Powered Hiring Tools to Identify and Reduce Biases in Candidate Selection. Path of Science, 11(4). https://doi.org/10.22178/pos.116-10
Summary. This article discusses how historical records, labels, and feature selection can create or preserve discriminatory outcomes in AI-supported recruitment. It proposes more transparent, accountable hiring processes centered on bias mitigation, human oversight, and regular auditing.
Key claims
- Historical hiring data can encode prior exclusion and therefore cannot be assumed to provide neutral training signals for recruitment models.
- Fairness risks can enter through training data, labels, features, and evaluation choices, not only through the final decision rule.
- The article recommends data and system audits, explainability, and human oversight as components of equitable AI-supported hiring.
Evidence & limitations
- The article is primarily a literature-based discussion and guideline proposal; it does not report a controlled validation of a specific hiring model or intervention.
- Its recommendations should be interpreted alongside stronger empirical and legal evidence on selection validity, adverse impact, and governance.