World Bank Pushes for 'Small AI' Running on Phones in Developing Countries

The World Bank has called on developing countries to build "small" artificial intelligence models that operate on mobile phones even with weak connectivity, according to a report by The Next Web citing The Guardian.

Pascal Donohoe, Chief Knowledge Officer at the World Bank, said that AI could accelerate growth in emerging economies, provided it meets local needs. He noted that small, specialized models could operate in local languages and on weak networks, enabling development that used to take centuries to occur within decades.

He cited the example of Indian farmers in the "AgriConnect" program who need soil analysis in their own language, stating that there will never be enough scientists and experts to serve every small farm, and that these tools could come jointly from the World Bank, governments, and private companies.

For her part, Kristalina Georgieva, Managing Director of the International Monetary Fund, said that governments must first secure electricity, digital infrastructure, and protection from bad actors, warning that 40% of jobs worldwide could be altered or lost due to AI, with young people being the most vulnerable.

What Do These Terms Mean?

Small Artificial Intelligence: Compact AI models trained on limited data for a single task, operating with fewer resources.

Specialized Model: An AI program trained on a single domain, such as soil analysis or translation, rather than general knowledge.

Digital Infrastructure: The networks, servers, and communications that enable online services and deliver AI to users.