We propose an extension to occupation-level measures of Artificial Intelligence (AI) exposure, accounting for AI's potential as either a complement or a substitute for labor, where complementarity reflects lower risks of job displacement. We validate the classification on US Commuting Zones (CZs) using a large dataset on online job postings. CZs with greater employment in industries with greater AI adoption have witnessed a relatively larger decline in the share of vacancies in high-exposure low-complementarity occupations since 2019. Using worker-level microdata from 2 Advanced Economies (AEs) and 4 Emerging Markets (EMs), we find that AEs feature higher employment shares in high-exposure occupations, roughly evenly distributed between high- and low-complementarity jobs. Within countries, common patterns emerge in AEs and EMs, showing unevenly distributed risks and opportunities. Women and highly educated workers face greater occupational exposure to AI, at both high and low complementarity. Higher-earning workers are more likely to be in occupations with high exposure but also high potential complementarity.