World Bank Urges Developing Countries to Build Small AI
The World Bank, the global development lender, is urging developing nations to build "small AI": specialised artificial intelligence models that run on a mobile phone with low connectivity. Paschal Donohoe, its chief knowledge officer, made the case as the bank's annual meetings run in Bangkok, the Guardian reported on 18 October. More than two-thirds of rural schools in sub-Saharan Africa lacked steady internet, a bank report found.
Small AI uses models trained on narrower datasets for a specific purpose. Donohoe cited Indian farmers in AgriConnect, a World Bank agricultural scheme, who needed soil analysis in their own language. There will not be enough scientists and agricultural experts to do that for every small farm in India, he said. Such tools could come from the bank, governments and the private sector working together, he suggested. Donohoe said the bank mobilised private capital through guarantees as the US and UK cut aid budgets.
Small AI is a proposal: Donohoe named no budget, product or timetable. It needs power, and the same report found almost a third of rural schools in sub-Saharan Africa lacked electricity. Kristalina Georgieva, head of the International Monetary Fund, warned that AI is likely to augment or eliminate 40 percent of jobs worldwide, hitting young job starters hardest.
The annual meetings in Bangkok continue this week, with the effect of AI on jobs and growth as a central theme. Any small AI tools would depend on the bank, governments and private partners agreeing on them; the Guardian report names no date for a first product.
The cards of this story
from the video, click to enlarge
The text presents a concrete idea, small AI on phones, backs it with some figures and an example, and openly states that no budget, product or timetable exists.
The text names a specific response to a real problem: specialised AI models that work on phones with weak connectivity, illustrated by the AgriConnect soil analysis example. It supports this with figures on rural school connectivity and electricity, and it is clear that the idea is only a proposal without budget or date. It also mentions the IMF's warning about jobs, which adds balance. The score would rise with evidence that small AI works somewhere, such as results from a pilot, or with details on costs, power supply and who would run the tools.
AI assessment by fixed criteria. How we score →
