How urgently do AI-exposed jobs need retraining?

As artificial intelligence (AI) rapidly transforms the American labor market, workers are wondering how worried they should be about their occupation’s vulnerability to new technologies. 

A multidisciplinary team of researchers has developed a new metric: the Reskilling Urgency Index (RUI), that can provide an idea of the uncertainties faced for people working in 879 occupations, based on factors that extend past the skills of the job itself.

While previous workforce studies focused almost exclusively on how heavily jobs are exposed to AI, the team extended that research by integrating two vital complementary indicators: skill non-transferability and job non-specialization. By adding the component of location, the resulting combined metric provides a framework for evaluating true workforce vulnerability.

In a paper titled "Mapping Workforce Reskilling Urgency in the Age of AI: Occupational and Geographic Patterns across the United States," published online in Science Direct, the researchers show that the additional metrics more accurately describe the potential labor disruption AI poses to individual occupations with a state.

Created by Waris Faizi, sociology Ph.D. student, and Junghwan Kim, assistant professor of geography, with collaboration from Dayoung Kim, assistant professor in the Department of Engineering Education; Can Dogan, associate professor of economics at Radford University; and Sohyun Park, assistant professor in computational and data science at George Mason University’s Korean campus, the index ranks 879 occupations based on how urgently workers need to acquire new skills to adapt within their current roles or transition to new ones.

"The index ranks occupations so people know where reskilling is required as soon as possible, whether to maintain their current roles or prepare for new ones in the job market," Faizi said. "If a job has high AI exposure, how necessary is it for those workers to be reskilled?”

Beyond exposure: The mechanics of reskilling urgency

Faizi conceived the project while taking a graduate course taught by Junghwan Kim on applying generative AI in social science research. To move the concept from a theoretical model into an actionable policy framework, the researchers expanded the collaboration to incorporate expertise in engineering education, workforce policy, and labor economics.

"Identifying occupations exposed to AI is only part of the workforce challenge," said Dayoung Kim, who brought experience in labor market data and state-level workforce policy to the project's conceptual and methodological design. "A more actionable question is: which workers face the most urgent need for reskilling when their occupations are exposed to AI, and where are those needs concentrated?"

By grounding the index in dataset metrics from O*NET and employment statistics from the U.S. Bureau of Labor Statistics, the researchers evaluated both individual occupational profiles and state-level workforce patterns. 

Occupations ranking highest on the index include:

•  Administrative services managers
•  Purchasing agents
•  Web developers
•  Paralegals 
•  Computer systems architects

The lowest-urgency occupations are partly manual labor roles and highly specialized professionals such as pilots, dentists, and radiologists, due to lower direct exposure to generative AI models. 

“Our key finding is an inverted-U pattern,” Faizi said. “Reskilling urgency is highest for occupations of intermediate specialization and lower at both ends — the least-specialized manual roles and the most-specialized professional roles.”

However, the researchers emphasize that the index should serve as a proactive planning roadmap not an alarm.

"Workers in high-RUI occupations do not need to rush into changing careers, but they should actively consider how their occupations are evolving, how AI can complement the work they currently perform, and which additional or transferable skills could make them more adaptable," Dogan said. "Ideally, reskilling should occur while people are still employed and have time to prepare, rather than after technological change has forced an occupational transition."

Geographic implications and workforce strategy

The researchers mapped occupational RUI scores against employment densities across the United States, creating a spatial distribution that demonstrates how reskilling urgency does not adhere to traditional regional boundaries. 

“The distribution of reskilling urgency is not geographically uniform,” said Junghwan Kim, the paper’s corresponding author. “States differ in their occupational employment composition, while occupations vary substantially in their RUI scores. Together, these differences produce distinct geographic patterns of reskilling urgency across the country, highlighting the need for place-based workforce development and reskilling strategies rather than a one-size-fits-all approach.”

Higher concentrations of reskilling urgency appear in Washington, D.C., alongside states such as Massachusetts, Maryland, Illinois, Texas, Arizona, Connecticut, Utah, and Florida — regions characterized by heavy concentrations of professional, technical, and administrative employment.

"Integrating occupation-level RUI scores with state employment patterns allowed us to move from identifying vulnerable occupations to identifying where reskilling needs may be especially concentrated," Dayoung Kim said. "This provides a more granular basis for workforce strategy: states with high employment-weighted RUI scores may benefit from more immediate reskilling interventions, while training strategies can be targeted toward workers whose existing skills provide fewer pathways into adjacent occupations."

Original study: DOI.10.1016/j.tele.2026.102445