E-Recruitment and Cultural Bias: A Structural Audit and Detection Design for LLMs

Digital recruitment platforms are often seen as neutral systems aimed at improving efficiency and fairness in hiring. However, new research shows that algorithmic and AI tools can mirror and intensify social and cultural biases. As Large Language Models (LLMs) become more common in recruitment tech, concerns grow about how cultural values in training data and model structures may influence hiring outputs. This study explores the existence and nature of cultural bias in AI-supported e-recruitment, challenging the idea of technological neutrality. It has two main goals: first, to analyze how bias arises not only from technical errors but also from deeper sociocultural elements embedded in digital platforms and decision-making frameworks; second, to develop and test a methodology for identifying and analyzing cultural bias in LLM-generated recruitment discussions. The approach involves a two-step process: a systematic review of 42 peer-reviewed articles (2010–2025) from Web of Science, offering an interdisciplinary view of bias in e-recruitment across management, sociology, and computer science; and a structural audit combining cross-cultural management theories with machine learning. Empirically, responses to 25 intercultural recruitment scenarios created by three LLMs from different geopolitical backgrounds are analyzed. Cultural bias is evaluated using the Covert Harms and Social Threats (CHAST) framework, which detects subtle communicative harms in areas like competence, morality, and opportunity. Results show that cultural bias in LLM outputs is systematic, not accidental, and varies across models. The differences are linked to the sociocultural contexts of model development, leading to ethnocentrism and stereotypical cultural framing. By merging cultural theory and computational analysis, the study advances the emerging field of cross-cultural management and AI ethics. It offers a replicable framework for diagnosing structural bias in AI-driven recruitment and stresses the importance of governance to ensure AI supports inclusive, fair hiring rather than reinforcing cultural hierarchies.