The World Bank is encouraging developing countries to invest in the creation of “small AI” models—artificial intelligence systems designed to operate on mobile devices with limited connectivity—arguing that such technology could significantly accelerate economic development. Paschal Donohoe, the World Bank’s chief knowledge officer, emphasized that while large tech firms often focus on sophisticated consumer applications, AI adapted to local contexts could offer substantial benefits for emerging economies.

Small AI refers to specialised models trained on narrower datasets for specific tasks, enabling them to function effectively in environments with limited infrastructure. Donohoe explained that these models could deliver essential services in local languages, directly on mobile phones, even where internet access is inconsistent or unavailable. He cited the example of Indian farmers participating in the World Bank’s AgriConnect agricultural programme, who need real-time soil analysis in their own languages—a demand that cannot be met solely by human experts. Small AI, he said, could fill this gap by providing accessible, targeted information at the grassroots level.

“The potential is there for economic development that might have taken centuries to occur to instead happen within decades,” Donohoe said, highlighting the transformative capacity of AI when tailored to the needs of low-resource settings.

The initiative promotes partnerships between the World Bank, governments, and the private sector to develop and deploy these localized AI tools. This focus comes amid discussions at the annual meetings of the World Bank and International Monetary Fund (IMF), held this week in Bangkok, where AI’s impact on jobs and economic growth has been a central topic.

IMF Managing Director Kristalina Georgieva underscored the foundational role of infrastructure and governance in enabling successful AI adoption. “You need access to electricity. You need access to digital infrastructure. People have to be geared to adoption. And the government should be able to give some safety, protect against bad actors so that adoption can work,” she said. Georgieva also compared the potential democratizing effect of properly supported AI adoption to that of the cellphone revolution.

However, challenges remain. The World Bank’s recent world development report noted that in sub-Saharan Africa, nearly one-third of rural schools lack electricity, and over two-thirds lack reliable internet access, limiting AI’s reach. Georgieva also cautioned about AI’s disruptive effects on the labor market, predicting that as many as 40% of jobs globally could be transformed or eliminated by AI over time. She highlighted the risk this poses to young workers entering the labor market, who may find traditional entry-level roles increasingly scarce.

Under the leadership of President Ajay Banga, the World Bank has prioritized mobilizing private-sector investment for emerging economies. Donohoe stated that the institution had secured $120 billion in private capital over the past year through financial tools such as guarantees, amid reductions in official aid budgets by countries including the United States and the United Kingdom. He emphasized the importance of connecting private capital flows with job creation and economic advancement in developing nations.