Author(s) :

Melanie Mar.

Volume/Issue :

Abstract :

The rapid adoption of artificial intelligence (AI) across industries has prompted increasing interest in understanding its geographic distribution and underlying causes. This study investigates the factors and spatial characteristics of AI adoption across U.S. states using recent data from the U.S. Census Bureau’s Business Trends and Outlook Survey (BTOS), combined with state-level economic and demographic indicators. Specifically, the analysis incorporates gross domestic product (GDP), population, and unemployment rate as explanatory variables. A comprehensive methodological framework is employed, including Global Moran’s I for spatial autocorrelation analysis, ordinary least squares (OLS) regression, and spatial lag models based on both contiguity and distance-based weighting schemes. The results indicate that AI adoption exhibits no statistically significant spatial autocorrelation, suggesting a lack of geographic clustering across states. Furthermore, regression analysis indicates that even though GDP, population, and unemployment are related to AI usage, none of these variables are statistically significant predictors at conventional levels. Spatial lag models yield consistent findings, with insignificant spatial autoregressive coefficients, confirming the absence of meaningful spatial dependence. Model performance comparisons demonstrate that incorporating spatial effects does not improve explanatory power. These findings suggest that AI adoption in the United States is largely distributed and influenced more by localized, non-spatial factors than by regional spillover effects. The study emphasizes the need for incorporating more variables and finer spatial resolution in future research to better capture the complexity of technology adoption.