Assessing the Risks of Unemployment in the Age of Artificial Intelligence

This research article examines the relationship between artificial intelligence (AI) job postings and unemployment across 14 advanced economies from 2015 to 2023. It utilizes inverse probability weighted regression adjustment (IPWRA) with covariate balancing propensity scores (CBPS) to estimate the average treatment effect (ATE). The findings indicate that the impact of AI labor demand on unemployment is contingent upon macroeconomic conditions. Specifically, during high-inflation periods, an increase in AI job postings correlates with lower unemployment rates, suggesting smoother labor market adjustments and productivity gains. Conversely, in low-inflation periods, the relationship weakens, and may even reverse, indicating that cost-saving measures dominate when economic demand is low. The paper contributes to the existing literature by providing a cross-country econometric assessment and integrating technological change into a broader business-cycle framework, highlighting the complex dynamics between AI adoption and labor market outcomes.