World Bank says ‘small AI’ could transform developing countries
News Source
•Sun, 18 Oct 2026 10:00:56 GMT
📰 What Happened
The World Bank is urging poorer countries to build "small AI" models instead of chasing the biggest systems. These models can run on a phone with a weak internet connection.
Paschal Donohoe, the bank's chief knowledge officer, spoke in Bangkok. He said development that once took centuries could now happen in decades.
Small AI uses narrow data sets for one job, in local languages. He cited Indian farmers in a World Bank farm scheme who needed soil tests explained in their own language.
🔍 The Backstory
Big tech firms are racing to build larger and more costly AI models. Most of that work happens in the United States and China, in English.
Smaller models are cheaper to run and can work offline. That matters where data plans are expensive and power is unreliable.
Donohoe is a former finance minister of Ireland. The World Bank lends money to governments for development projects and advises them on policy.
🎯 Why It Matters
AI that works offline in a local language could help farmers and clinics far from big cities. It could also keep poorer countries from falling further behind.
The World Bank is urging poorer countries to build "small AI" models instead of chasing the biggest systems. These models can run on a phone with a weak internet connection.
Paschal Donohoe, the bank's chief knowledge officer, spoke in Bangkok. He said development that once took centuries could now happen in decades.
Small AI uses narrow data sets for one job, in local languages. He cited Indian farmers in a World Bank farm scheme who needed soil tests explained in their own language.
Big tech firms are racing to build larger and more costly AI models. Most of that work happens in the United States and China, in English.
Smaller models are cheaper to run and can work offline. That matters where data plans are expensive and power is unreliable.
Donohoe is a former finance minister of Ireland. The World Bank lends money to governments for development projects and advises them on policy.
AI that works offline in a local language could help farmers and clinics far from big cities. It could also keep poorer countries from falling further behind.