Author(s) :

Seohyun Hong.

Volume/Issue :

Abstract :

This study sought to examine how national diffusion of artificial intelligence measured by the share of artificial intelligence job postings is related to energy intensity and income. The hypothesis in this study was that the greater the diffusion of artificial intelligence is, the lower energy intensity is, reflecting efficiency gains, and higher income, reflecting productivity gains. Using two publicly available datasets, this study, this study merged country-year records, analyzing the most recent year with broad coverage in cross-sectional design. Outcomes identified energy use per unit of economic output and income per person by applying artificial job posting share as a single explanatory variable. Models were estimated with bivariate ordinary lease squares for the purpose of replication. Results indicated that there was a positive but statistically inconclusive association between the diffusion of artificial intelligence and primary-energy intensity. In addition, when focusing on electricity intensity, there was a small and statistically null association, while there was a positive and precise association between the diffusion of artificial intelligence and income as confirmed by three light robustness checks, including a logarithmic specification and a leave-one-out influence analysis. The findings in this study were not casual but descriptive. Future work is recommended to focus on extending to a short panel with fixed effects, emphasizing outcomes specific to electricity and carbon to add readily available controls and examine heterogeneity by region or income group.