![]()
How will artificial intelligence (AI) affect unemployment rates? The answer will vary depending on regional differences in occupation and industry—with urban technology hubs likely to see the greatest impact, reports a study from Taiwan in the June issue of the
Journal of Occupational and Environmental Medicine.
"Without intervention, unemployment in high-exposure cities could rise sharply under rapid AI adoption," according to the report by John Tayu Lee, PhD, of National Taiwan University, Taipei, and colleagues. They used measures of exposure to AI by industry and occupation to project trends in unemployment in Taiwan—a global leader in the technology industry—over the coming ten years.
The results suggested that occupational exposure to AI is "unevenly distributed" across regions. Industries such as finance, technology, and education had the highest AI exposure, while exposure was low for industries such as agriculture, construction, and accommodations and catering.
Urban centers in northern Taiwan with high concentrations of AI-intensive industries and occupations were projected to have the greatest impact on unemployment. In those cities, projected increases in unemployment by 2035 ranged from 16 percentage points in a scenario of moderate AI impact to 48 percentage points in a high-impact scenario. In contrast, southern and central regions with lower AI exposure were projected to see little or no impact on unemployment.
Younger workers aged 25 to 44 with university degrees were at the greatest risk of AI-driven disruption, particularly in AI-intensive industries. Occupational groups at high risk included professionals, managers, and office support staff.
The regional variations illustrate "the paradox of digital economies," according to the authors: "[R]egions that benefit most from innovation are simultaneously the most vulnerable to displacement." While rural areas may see little immediate impact of AI, they "risk long-term exclusion if they remain outside digital transformation pathways." The researchers point out some limitations of their analysis, emphasizing that the projections are "hypothetical simulations, not empirical forecasts."
Dr. Lee and coauthors discuss the need for "targeted workforce adaptation policies" such as skills retraining, social protection, and targeted innovation support to lessen the impact of rapid AI implementation on unemployment and regional disparities. The researchers conclude, "Only through such commitments can AI become a tool not just of efficiency, but of shared and sustainable progress."
Read the fully study here:
https://journals.lww.com/joem/fulltext/2026/06000/artificial_intelligence_and_heterogeneous.3.aspx