Algorithmic bias in HR recruitment systems: A qualitative analysis of managerial risk and sociological implications

Baijia Song

PLOS ONE, 2026, vol. 21, issue 6, 1-31

Abstract: Algorithmic hiring systems have rapidly proliferated across industries, promising improved efficiency, objectivity, and scalability in recruitment processes. However, growing empirical evidence reveals a gap between these expected benefits and actual outcomes, as many systems inadvertently reproduce or amplify historical inequalities embedded in training data. The main aim of this study is to develop and evaluate a rigorous, multi-layered framework capable of identifying, interpreting, and mitigating bias throughout the full lifecycle of algorithmic hiring systems, ensuring both immediate decision fairness and long-term career equity. To achieve this aim, the Multi-Layer Bias Analysis and Mitigation System (ML-BAMS) is introduced as a comprehensive approach to detecting and mitigating bias in HR recruitment systems. The framework integrates modules for bias decomposition and data diagnostics, recruitment-aware fairness evaluation across multi-stage pipelines, interpretability of opaque models through influence mapping, fairness-preserving representation learning, and longitudinal simulation of career mobility outcomes. Using synthetic hiring datasets, the proposed framework demonstrates substantial reductions in demographic disparities across key fairness metrics, including improvements in demographic parity (32.4%), equal opportunity (28.7%), equalized odds (25.9%), and treatment equality (19.2%), while maintaining competitive predictive accuracy (ΔAccuracy: + 0.024). These findings highlight the importance of integrated sociotechnical approaches that address bias transmission, enhance transparency, and account for long-term impacts. The ML-BAMS framework provides a practical and modular toolset for implementing responsible AI in recruitment, balancing operational performance with ethical considerations of fairness and social equity.

Date: 2026
References: Add references at CitEc
Citations:

Downloads: (external link)
https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0349400 (text/html)
https://journals.plos.org/plosone/article/file?id= ... 49400&type=printable (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:plo:pone00:0349400

DOI: 10.1371/journal.pone.0349400

Access Statistics for this article

More articles in PLOS ONE from Public Library of Science
Bibliographic data for series maintained by plosone ().