Abstract

This research investigates the pervasive issue of algorithmic bias in machine learning (ML) models used for talent acquisition. As organizations increasingly rely on automated systems to screen resumes, identify qualified candidates, and even conduct initial interviews, the potential for perpetuating and amplifying existing societal biases becomes a significant concern. This paper presents a comparative analysis of several commonly used ML models in recruitment, evaluating their performance across different demographic groups. It identifies sources of bias within these models, stemming from both data and algorithmic design. Furthermore, it explores and evaluates various mitigation strategies, including data pre-processing techniques, algorithmic adjustments, and post-processing interventions, aimed at enhancing fairness and promoting diversity and inclusion in the hiring process. The findings highlight the importance of careful model selection, robust bias detection, and proactive implementation of mitigation strategies to ensure equitable talent acquisition practices. The study contributes to the growing body of knowledge on responsible AI in HR and offers practical recommendations for organizations seeking to leverage ML for talent acquisition while upholding ethical principles.