AI and the Ladder of Development
Working paper · Theory · AI and development
Abstract
Will artificial intelligence help developing countries ascend the development ladder, or remove some of its rungs? I develop a small open-economy model that allocates labour across manufacturing, progressive tradable services, and local non-tradable services. Productive capability governs both absorptive capacity and long-run learning.
Frontier AI raises domestic productivity in tradable sectors but also intensifies foreign competition. Its effect on capability growth decomposes into a positive covariance term—whether AI strengthens sectors with the greatest learning spillovers—and a negative reallocation term generated as higher wages move labour toward weak-learning local services. AI can therefore raise average productivity while lowering long-run capability growth, producing a potential development trap.
Research themes
Artificial intelligence · Structural transformation · Learning spillovers · Open-economy growth
Current draft available on request — t.dobermann@lse.ac.uk