Convergence analysis of artificial intelligence research capacity: Are the less developed catching up with the developed ones?

Abstract

This study examines whether less developed countries are catching up with developed ones using the log t convergence technique (LCT) and the dynamic spatial ordered probit (DSOP) model. The findings revealed that first, there is no overall convergence in AI research capacity. Second, club clustering analysis showed convergence in four of the five groups of countries on AI research output and in three of the four groups on AI patent grants. Third, the countries are experiencing a slow divergence process in AI research capacity. Fourth, the region, income group and cluster of the countries are influencing the convergence process.