Large Language Models in human resource management: A systematic literature review of applications, open issues and future research directions

Rapid adoption of large language models (LLM) in human resource management (HRM) is transforming core functions that include recruitment and selection, training and development, and performance evaluation. Although prior work offers conceptual insights, empirical evaluations, and a methodological debate on AI in HRM, there is no unified synthesis of LLM-specific contributions. In this paper, we present a systematic review of the literature on LLM-enabled HRM, integrating both conceptual frameworks and empirical findings. In particular, we categorize studies by LLM architecture (discriminative vs. generative) and functional HRM application (e.g., automated screening, sentiment analysis for performance reviews). We also examine ethical considerations-including bias mitigation, data privacy and transparency and we further assess organizational impact, such as workflow integration and user acceptance. By mapping the current state of knowledge, we identify open issues around models reliability, explainability, reasoning capabilities and the pertinence of human workforce in the agentic era. Finally, we outline future research directions to advance evidence-based practice in the era of AI-enhanced workforce management.