AI exposure and task detachability: migrant anchoring and work-linked health protection among highly educated foreign-born adults in the United States
Year: 2026
Type: article
Study design: Observational Study
Abstract: Background: Occupation-level AI exposure indices describe the potential overlap between occupational work content and AI capabilities rather than individual AI use or employer adoption. However, task organisation and access to work-linked protections may differ across occupational contexts. We examined whether associations between a compound high-AIOE/high-task-detachability occupational profile and upstream insurance and material conditions varied by migrant anchoring among highly educated foreign-born adults in the United States. Methods: We linked 2017-2024 IPUMS American Community Survey (ACS) data for adults aged 25-64 with a bachelor's degree or higher to AI Occupational Exposure (AIOE) and O*NET 30.3 measures. High AIOE was defined as the top occupational quartile. The focal profile combined high AIOE with high task detachability and was compared with lower-AIOE occupations. Migrant anchoring combined citizenship, US residence duration, English proficiency, family co-residence, and homeownership; scores of 0-2 indicated lower anchoring. The three principal domain-specific outcomes were absence of employer- or union-linked health coverage from any eligible source among employed wage and salary workers, uninsurance, and family income below 200% of the federal poverty threshold. All-adult work-linked Social Reproduction Instability (W-SRI) ≥ 2 was retained as the focal cumulative summary measure. Weighted models were adjusted for demographic, regional, annual, and broad occupational characteristics. Inference used CR2/Satterthwaite covariance, wild cluster bootstrap tests, and leave-one-cluster-out analyses. Results: The analysis included 746,995 foreign-born person-year observations, of which 22.47% were classified as having lower anchoring. Among adults with higher anchoring, the focal occupational profile and lower-AIOE occupations showed few differences in the four principal outcomes. Among adults with lower anchoring, the focal occupational profile was associated with a lower prevalence of all four adverse outcomes. The corresponding between-stratum differences were -10.68 percentage points for absence of employer- or union-linked coverage, -6.75 percentage points for uninsurance, -7.90 percentage points for poverty, and -7.55 percentage points for cumulative burden. Magnitudes remained similar after excluding 2020, restricting the sample to 2018-2024, adding state fixed effects, and restricting the sample to currently employed respondents. In the labour force-restricted analysis, the unemployment modification was -0.37 percentage points (95% CI -1.15 to +0.42) and was imprecisely estimated. Conclusion: Among highly educated foreign-born adults with lower anchoring, the compound occupational profile was associated with a lower prevalence of coverage absence, uninsurance, poverty, and cumulative burden than occupations with lower AIOE. The findings describe occupation-level differences in upstream insurance and material conditions rather than causal effects of AI adoption or task detachability.
Source: Frontiers in Public Health
Authors: Tse Hou Hew, Rajah Rasiah, Nina Xie
Institutions: University of Malaya, Lingnan University
Language: English
Field: Medicine
Domain: Health Sciences
Open Access status: gold
Estimated APC paid: $3,801
OpenAlex ID: W7219647706