Artificial intelligence exposure and labor market transformations in Latin America

This paper examines the implications of artificial intelligence (AI) for employment, wages, and inequality in Latin America and the Caribbean (LAC). Using individual-level data from the STEP and PIAAC surveys and AI exposure indices from Webb (2020) and Felten et al. (2021), we estimate predicted AI exposure for workers across seven LAC economies (Bolivia, Chile, Colombia, Ecuador, El Salvador, Mexico, and Peru) and a set of OECD comparators. We then link occupation-level exposure to changes in employment and wages obtained from harmonized household surveys for the five LAC countries with comparable data over the period (Bolivia, Chile, Ecuador, Mexico, and Peru). The empirical strategy follows an expectation–maximization procedure that recovers individual-level exposure from occupation-based indices. The paper documents how AI exposure varies by skills, education, gender, and age, compares LAC patterns with those observed in OECD countries, and assesses distributional correlates across wage quintiles.