Not yet human-reviewed
This source record was written automatically by the research pipeline and has not yet been reviewed by the maintainer or integrated into a topic page. Treat it as a provisional summary and verify claims against the original source.
Fairness, AI & recruitment
Citation. Rigotti, C. & Fosch-Villaronga, E. (2024). Fairness, AI & recruitment. Computer Law & Security Review, 53, 105966. https://doi.org/10.1016/j.clsr.2024.105966
Summary. This peer-reviewed scoping literature review examines how fairness is defined, categorized, and implemented in AI-supported recruitment and selection. It argues that fairness cannot be reduced to a single technical metric because legal, ethical, and stakeholder perspectives differ, and it recommends cross-disciplinary and participatory approaches that prioritize applicants’ vulnerability in the hiring relationship.
Key claims
- AI recruitment tools can reproduce or perpetuate discrimination through data and decision processes, creating risks for applicants’ dignity, autonomy, well-being, and equitable labor-market opportunity.
- EU anti-discrimination law centers the treatment of applicants in recruitment processes, supporting a procedural account of fairness rather than evaluating only hiring outcomes.
- Legal compliance cannot be translated into one fairness metric or code threshold because anti-discrimination law serves multiple normative functions and requires contextual reasoning.
- The authors recommend cross-disciplinary and cross-sector collaboration, participatory stakeholder engagement, transparency about data and bias mitigation, and applicant-centered safeguards for fair AI hiring.
Evidence & limitations
- Scoping literature review spanning legal, ethical, and technical scholarship; it synthesizes definitions and recommendations rather than estimating causal effects or testing a specific hiring system.
- Published open-access article under CC BY 4.0; the original Jina markdown artifact was a DOI 404 stub, so the verified Leiden University repository publisher-version PDF replaced it. The Drive file hash above identifies the replacement PDF, not the discarded stub.