Does AI vibrancy reshape employment structures and unemployment? Short-run evidence from global panel data

This study examines whether national AI vibrancy affects employment structures and unemployment, utilising panel data for 36 countries (2017–2023) from the Stanford AI Vibrancy Index and World Bank indicators. Fixed- and random-effects models, along with Hausman tests and clustered standard errors, were applied. Results reveal no significant short-term effect of AI vibrancy on industrial or service employment, nor on unemployment. GDP per capita showed a modest positive effect on industry (β = 0.046, p < 0.10) and, in some models, on services (β = 0.045, p ≈ 0.06). Labour force participation correlated positively with service employment (β = 0.322, p < 0.10) and negatively with unemployment (β = –0.006, p < 0.01). Strong time effects reflect pandemic-related labour shifts: industrial jobs declined, service employment grew, and unemployment spiked temporarily. Overall, labour dynamics appear driven more by global shocks and demographics than by AI vibrancy.