Artificial intelligence diffusion, adaptive Human-Capital Transformation, and unequal development: evidence from a 50-country panel
Year: 2026
Type: article
Abstract: Introduction Despite expanding artificial intelligence (AI) research, evidence remains fragmented on how AI capacity, Human-Capital Transformation, institutions, and labor-market structure jointly relate to unequal development. Methods Using 229 country-year observations from 50 purposively selected economies over 2020–2024, this study integrates conceptual development, composite measurement, panel estimation, country configurations, and conditional projections for 2026–2027. Moving beyond traditional productivist paradigms, this study conceptualizes adaptive human capital (AHC) as the dynamic capacity to preserve, renew, recombine, and redeploy human capabilities under technological change, with health sustaining adaptability, lifelong learning enabling renewal, and skills recalibration supporting resilience. This model distinguishes adaptive capacity from its observable educational and digital foundations. Five author-constructed indices capture AI Diffusion Capacity (AIDC), Human-Capital Transformation (HCT), Enabling Conditions, AI-Related Labor-Market Pressure, and multidimensional Development Outcomes. Two-way fixed-effects models use country-clustered standard errors, restricted wild-cluster-bootstrap inference, and false-discovery-rate correction. Sensitivity checks examine alternative weights, component exclusion, and 1,000 positive random-weight draws. Results AIDC coefficients are positive for HCT (β = 0.220, p = 0.0063) and Development Outcomes (β = 0.160, p = 0.027), although the latter attenuates after income adjustment. The strongest results concern labor absorption: negative AIDC coefficients for total unemployment (β = −0.359) and youth unemployment (β = −0.319) and a positive coefficient for the employment-to-population ratio (β = 0.258) remain significant after income adjustment and false-discovery-rate correction. Comparisons among similar AIDC countries identify configuration-level heterogeneity without a consistently dominant human-capital, institutional, or labor-market correlate. Complementary AIDC–HCT and AIDC–unemployment maps distinguish aligned and asymmetric country configurations; the latter classifies 14, 12, 11, and 13 economies into four labor-market groups. Conditional joint AIDC–HCT paths imply declining Development Outcomes under the lower scenario and improving outcomes under the central and upper scenarios; separate AIDC paths map labor-configuration transitions. Discussion Unequal development is reflected in asymmetric technological and labor-market configurations: 12 countries combine above-average AIDC with above-median unemployment, whereas 11 combine below-average AIDC with unemployment at or below the sample median of 4.545%. The study connects an adaptive human-capital framework with differentiated empirical evidence on technological capacity, human capabilities, and labor absorption, informing country-specific priorities for technological scaling, capability renewal, and support during labor-market transitions.
Institutions: Escola de Administração de Empresas de São Paulo, Fundação Getulio Vargas, Moscow State University of Psychology & Education, Tashkent State University of Economics, Plekhanov Russian University of Economics, Financial University, Moscow Polytechnic University, North-Eastern Federal University
Language: English
Subfield: Economics and Econometrics
Domain: Social Sciences
Open Access status: gold
Estimated APC paid: $2,353
OpenAlex ID: W7225835615