Dynamic Identification of Generative AI Skills: A Structural Topic Modeling Approach
提出一种结构主题建模框架，利用实时招聘数据动态识别生成式AI技能需求，并估计其劳动力市场溢价，为工程管理者提供决策支持。
The accelerated adoption of generative artificial intelligence (AI) is transforming engineering management practices, driving rapid changes in labor market demand for generative AI skills (GAIS). We propose a novel structural topic modeling (STM) framework for the dynamic identification of in-demand GAIS using free-form job description texts. We use STM with real-time job postings data to discover in-demand GAIS and estimate the labor market premium associated with these skills. We illustrate the proposed framework using a large corpus of generative AI job postings to identify the ten most salient GAIS. Further, we show how skill identification can be conditioned on specific job characteristics, such as required experience or job location. Finally, we demonstrate how the STM framework can be used as a decision-support system, providing practical insights into various real-world decisions faced by engineering management professionals and organizations. Theoretically, our study contributes to the engineering management literature by operationalizing a dynamic capabilities approach to workforce agility. Practically, it provides an evidence-based tool for engineering managers to guide workforce development and hiring strategies using real-time labor market insights. More generally, our study offers a practical framework for engineering managers to continuously sense and respond to evolving labor market demands in fast-changing technological domains.