Assessing the Limits of Algorithms: Reorienting Human Resource Management Policies for Mitigating Bias in AI-Based Performance Evaluation

Abstract

This study aims to analyze the urgency of reorienting Human Resource Management (HRM) policies and updating operational procedures to mitigate the risks of bias generated by Artificial Intelligence (AI)-based performance evaluation systems. Employing a qualitative-descriptive approach, this study combines a literature review and field research through in-depth interviews to examine regulatory frameworks, organizational internal policies, and case studies concerning the implementation of performance evaluation algorithms in the digital industry sector. The findings reveal that the use of AI in performance evaluation systems has significant potential to generate historical data bias, algorithmic unfairness toward minority groups, and the erosion of human empathy in competency assessments. Therefore, the reorientation of HRM policies is imperative through the formulation of multilayered supervisory mechanisms (Human-in-the-Loop), periodic algorithm audits, and transparency in performance evaluation metrics for employees. Theoretically, this study contributes to enriching the body of knowledge in strategic human resource management, particularly in the area of technology governance in the era of digital disruption. Practically, it provides strategic guidance and concrete policy recommendations for HR practitioners and policymakers in designing performance evaluation standard operating procedures (SOPs) that are not only technologically efficient but also uphold the principles of fairness, accountability, and organizational humanism.