What question did this study set out to answer?

This research examines how governance models affect bias in algorithmic hiring practices during early-stage recruitment.

June 13, 2026Open Access

Making Bias Visible: Algorithmic Hiring and the Role of Governance in Early-Stage Recruitment

Key Points

This research examines how governance models affect bias in algorithmic hiring practices during early-stage recruitment.
Comparative analysis of New York City's Local Law 144 and North Carolina's Responsible AI Framework.
Assessment of audit documentation and regulatory guidance.
Evaluation of the impact of mandatory versus voluntary governance on AI hiring practices.
Limited oversight in voluntary frameworks obscures algorithmic decision-making.
Significant gaps noted in transparency and accountability measures in AI hiring.
Findings indicate that governance frameworks directly influence economic opportunity access.

Abstract

Artificial intelligence (AI) increasingly shapes early-stage recruitment, often determining which candidates advance before human review. Although often framed as efficient and objective, these systems can reproduce or amplify labor-market inequalities by generating disparate impacts--outcome-based differences in selection rates across protected groups, regardless of intent (Barocas & Selbst, 2016). Using a comparative approach, it analyzes New York City’s Local Law 144, which mandates bias audits for Automated Employment Decision Tools, and North Carolina’s voluntary Responsible AI Framework. Drawing on policy analysis, regulatory guidance, and audit documentation, the study assesses the influence of mandatory versus voluntary governance models on transparency, accountability, and equity. Findings reveal significant gaps, showing that limited oversight can obscure algorithmic decision-making and weaken bias mitigation. By focusing on entry-level recruitment, this research highlights how governance frameworks shape access to economic opportunity and offers insights for more equitable AI hiring practices.

Demander à l'IA

Bookmark

View Full Paper