Abstract
This study investigates the near-future impacts of generative artificial
intelligence (AI) technologies on occupational competencies across the U.S.
federal workforce. We develop a multi-stage Retrieval-Augmented Generation
system to leverage large language models for predictive AI modeling that
projects shifts in required competencies and to identify vulnerable occupations
on a knowledge-by-skill-by-ability basis across the federal government
workforce. This study highlights policy recommendations essential for workforce
planning in the era of AI. We integrate several sources of detailed data on
occupational requirements across the federal government from both centralized
and decentralized human resource sources, including from the U.S. Office of
Personnel Management (OPM) and various federal agencies. While our preliminary
findings suggest some significant shifts in required competencies and potential
vulnerability of certain roles to AI-driven changes, we provide nuanced
insights that support arguments against abrupt or generic approaches to
strategic human capital planning around the development of generative AI. The
study aims to inform strategic workforce planning and policy development within
federal agencies and demonstrates how this approach can be replicated across
other large employment institutions and labor markets.