Research Summary: Complementarity, Augmentation, or Substitutivity? The Impact of Generative Artificial Intelligence on the U.S. Federal Workforce

A research study evaluating several hundred U.S. federal occupational series using a multi-stage Retrieval-Augmented Generation (RAG) framework demonstrates that generative artificial intelligence primarily complements and augments public sector roles rather than substituting them. The analysis indicates that even in occupations with relatively high substitutivity metrics, the observed impact is confined to the partial automation of routine subtasks rather than widespread workforce or position elimination.

Why it matters

This analysis clarifies that near-term generative AI deployment in government should be treated as a workforce enhancement and transformation initiative rather than an exercise in headcount reduction. Strategic decision-makers can deprioritise broad workforce rationalisation fears and focus instead on targeted capability evolution. It provides empirical grounding for public sector modernization, indicating that productivity gains will come from task redesign rather than occupational displacement.

Key insights

• Generative AI predominantly acts as a complementary and augmenting capability for U.S. federal occupational series, showing minimal capacity for outright job substitution.
• The analytical framework scored occupational series across three dimensions: complementarity (enhancing competencies), augmentation (requiring transformed competencies), and substitutivity (replacing tasks).
• Occupations exhibiting the highest substitutivity scores only undergo partial automation of routine subtasks, not full role or position elimination.
• The methodology utilised a multi-stage Retrieval-Augmented Generation system to ground LLM-based workforce evaluations directly in curated evidence.

Source

arXiv — Computers and Society — https://arxiv.org/abs/2503.09637