AI and the Global Productivity Divide: Fuel for the Fast or a Lift for the Laggards?

Artificial Intelligence (AI) has the potential to be an important driver of productivity growth over the next decade, even if with significant cross-country heterogeneity. This paper examines the potential of AI to foster productivity growth in Low-Income Countries (LICs) and Lower-Middle-Income Countries (LMICs). LICs and LMICs risk benefiting less from AI due to low incidence of knowledge-intensive services, where gains from AI mostly occur. Additionally, barriers to AI adoption include inadequate digital infrastructure, low levels of education and skills in the workforce, limited access to financing for high AI adoption costs, and underdeveloped regulatory frameworks. At the same time, LICs and LMICs may benefit from factors such as a young workforce and international spillovers through knowledge transfers. Overall, structural weaknesses in LICs and LMICs risk outweighing these potential advantages. This underscores the need for policies that enhance capabilities for AI adoption in LICs and LMICs and help seizing long-run opportunities from the global AI economy.