small AI
In plain wordsCheap, compact AI tools built for one practical job — like advising farmers or screening patients — that poorer countries can afford to run, instead of giant all-purpose models.
Compact, low-compute AI models and applications designed for specific practical tasks — crop and weather advice, disease screening, tutoring, voice services on basic phones — that are affordable to build, run and adapt in resource-constrained settings. The World Bank made it a centerpiece of its advice to developing countries in 2026, recommending that they adopt existing open-source or commercial models, adapt them to local languages, climates and health protocols, and invest in electricity, connectivity, data and skills rather than try to build frontier foundation modelsfoundation model, n.: A very large AI model, built once at great cost from vast amounts of text, images or code gathered mostly from the internet, that many products then reuse as their starting point, from chatbots to coding tools., which it judged cost-prohibitive given the concentration of advanced compute in the United States and China. Examples it cited include AI-based weather forecasts for smallholder farmers in India and a low-cost math-tutoring platform in Ghana.
The ministry chose small AI: a voice advisory bot in three local languages, running on a modest cloud budget.
Origin
Promoted by the World Bank, notably in its World Development Report 2026 on artificial intelligence, released in early August 2026, which urged developing economies to adopt and adapt existing AI rather than chase large frontier models.
Usage note
Not a technical category like small language modelsmall language model, n.: A language model built to be compact enough to run cheaply and quickly, even on a laptop or phone, trading some knowledge and skill for speed, privacy and lower cost.: the emphasis is on cost and fitness for local purpose, whatever the architecture.
Related
- small language modelsmall language model, n.: A language model built to be compact enough to run cheaply and quickly, even on a laptop or phone, trading some knowledge and skill for speed, privacy and lower cost.
- edge AIedge AI, n.: Doing AI's computing on or near the devices that collect the data, such as cameras, cars, factory machines, phones or local servers, instead of sending everything off to distant data centers.
- on-device AIon-device AI, n.: AI that runs directly on your own phone, laptop, car or headset rather than on a company's distant servers, so it can work offline, respond instantly and keep your data on the device.
- digital dividedigital divide, n.: The gap between people, places or countries that have good access to digital technology, and the skills to use it, and those that do not; AI may widen it.
- sovereign AIsovereign AI, n.: A country's ability to build and run AI on its own terms, so that the computers, models and know-how it relies on cannot be withheld or dictated by foreign companies or governments.
- AI for goodAI for good, n.: Using AI to tackle social and environmental problems, such as disease, poverty, disasters or climate change, rather than only to make a profit.
- frontier modelfrontier model, n.: One of the most capable AI models in existence at a given time, built by a handful of leading labs at great cost, and the main target of AI safety rules.
Sources
- World Bank (2026). World Development Report 2026. Washington, DC: World Bank. worldbank.org/en/publication/wdr2026
- Engineering News (2026). World Bank proposes small, low-cost AI for developing countries. 4 August 2026. engineeringnews.co.za/article/world-bank-proposes-small-low-cost-ai-for-developing-countries-2026-08-04