World Bank Backs Small AI for Developing Economies

The World Bank is urging developing economies to build “small AI” systems designed to operate on mobile phones, work with limited connectivity and address local needs. Speaking to The Guardian during the IMF–World Bank Annual Meetings in Bangkok, Paschal Donohoe, the World Bank Group’s managing director and chief knowledge officer, argued that locally adapted AI could help emerging economies accelerate development. The idea is to make useful AI available beyond countries and companies with access to expensive computing infrastructure. The Guardian

Donohoe said economic development that might otherwise take centuries could potentially happen within decades. This is an optimistic assessment of AI’s potential, not a measured forecast or a guaranteed outcome. The broader discussion reflects a shift towards practical applications that improve productivity and public services, rather than focusing exclusively on increasingly powerful general-purpose AI models. 

India is an important part of the discussion. In its October 6, 2026, India Development Update, the World Bank identified small AI as one way to adapt technology to local needs. The report says the World Bank Group is investing in targeted, locally adapted tools through AgriConnect, with the aim of making AI’s benefits more accessible. It also emphasises the need to strengthen infrastructure, skills and the business environment so that adoption can translate into broader economic gains.  World Bank 

The agricultural use case is straightforward: India has a large farming population, and specialised digital tools could help farmers access information such as soil analysis in their own languages. Donohoe told The Guardian that AI could help address the shortage of scientists and agricultural experts available to support every small farm. The opportunity is to extend access to useful knowledge, not to assume that an AI model can replace expert testing, local agricultural extension services or human judgement. 

The World Bank’s argument is also connected to employment. At the Bangkok meetings, leaders discussed how AI could improve productivity and support job creation in emerging economies. For countries with large young populations, locally relevant applications could help workers, farmers and small businesses use technology without having to build frontier-scale AI infrastructure themselves. The result will depend on whether these tools are affordable, useful and supported by appropriate skills and institutions.  

There is also a financing question. Donohoe told The Guardian that the World Bank had mobilised $120 billion in private capital over the preceding year through guarantees and other financial instruments. That figure refers to private capital mobilisation across the bank’s work; it should not be interpreted as a dedicated $120 billion fund for small AI. The reporting does not establish a separate small-AI funding allocation or a universal rollout timetable. The Guardian