World Bank pushes low-bandwidth AI models to boost growth in poor nations

During the Bangkok meetings of the World Bank and IMF, the Bank called on low-income countries to develop “small AI” models that function on mobile devices with limited connectivity. Paschal Donohoe explained that such specialized models, trained on focused data, can provide services such as soil-analysis translations for Indian farmers participating in the AgriConnect program. He suggested partnerships among the Bank, governments and private firms to create these tools.

IMF chief Kristalina Georgieva added that reliable power, internet access and protective regulations are essential for democratic AI adoption. The Bank’s recent development report noted that many rural schools in sub-Saharan Africa still lack electricity and stable internet, underscoring the infrastructure gap. Georgieva warned that AI could affect up to 40% of jobs, especially for young workers entering the labor market. Under President Ajay Banga, the Bank has mobilised $120bn of private capital to support such initiatives.