Integrating Machine Learning in HR: Ethical Dilemmas and Balancing Efficiency with Fairness

This paper proposes a new deep learning approach that is an amalgamation of both models as a solution to the ethical challenges in HRM with a focus on intersectional fairness and pay gap issues. Combining XAI and the fairness metrics, the paper explains how the application of ML in HR decision-making can optimize efficiency and interpretability. The proposed model, estimated on a set of 50 K employee records at the company, enhances traditional ML methods in terms of providing fairness for gender, race, and age intersections. They reveal higher average positive impacts on both predictive performance and fairness, particularly in hiring or promoting employees and determining their wages. Therefore, the paper stresses openness through explainability and outlines the future direction in the integration of AI and human supervision for ethical HRM.