Identifying skill requirements for emerging occupations using large language models and recruitment data

Abstract

The rapid advancement of digital technologies and the transformation of labor markets have led to the emergence of new occupations, and there is no redefinition of the emerging job titles. Understanding the evolving skillsets required for these occupations is critical for aligning workforce training, higher education, and policy development with industry needs. This study explores the application of artificial intelligence, specifically OpenAI, to extract and analyze soft and hard skills from a recruitment dataset in order to understand Vietnamese labor market trends. Using LLM, the research identifies the most frequent job titles and skill requirements across industries. The results reveal that Manager, Specialist, and Assistant are the most common job titles, suggesting strong demand for supervisory, expert, and support-level positions. Other frequently mentioned roles include Executive, Engineer, Director, Analyst, and Developer, while Intern and Trainee indicate opportunities for early-career professionals. By mapping these evolving skill clusters, the research contributes a data-driven framework for detecting labor market shifts in real time. The implications extend to higher education institutions seeking curriculum innovation, employers designing recruitment strategies, and policymakers aiming to predict the trend of the labor market in the era of Industry 4.0.