This study presents a global analysis of the potential effects of generative AI on employment. Using the GPT-4 model, we estimate task-level exposure scores and assess their potential employment impacts globally and across country income groups. We find that clerical work is the only broad occupational category highly exposed to the technology, while other occupational groups such as managers, professionals and associate professionals exhibit much lower exposure levels. Consequently, the primary impact of generative AI is likely to be the augmentation of work rather than the full automation of occupations. Due to different occupational structures, employment effects vary across countries. In low-income countries, only 0.4 percent of total employment is potentially exposed to automation effects, compared with 5.5 percent in high-income countries. The effects are also highly gendered, with women more than twice as likely as men to be affected by automation. We find that 10.4 percent of employment in low-income countries has the potential to be augmented, compared with 13.4 percent in high-income countries. However, these estimates do not consider infrastructure constraints, which may significantly limit adoption in lower-income contexts.