The World Bank has issued a new recommendation for emerging economies to focus on the development of 'small AI' rather than attempting to replicate the massive, resource-heavy models favored by Silicon Valley. According to Paschal Donohoe, the World Bank’s chief knowledge officer, these smaller models are designed to operate on standard mobile phones in areas with poor internet connectivity, making them more accessible and relevant to local needs.

Unlike the broad, general-purpose AI systems being developed by major tech firms, 'small AI' relies on specialized models trained on narrower, localized datasets. The World Bank suggests that this approach allows developing countries to address specific economic challenges more efficiently. By tailoring technology to local requirements, the organization believes these nations can bypass the high infrastructure costs associated with large-scale artificial intelligence, which often requires significant computing power and high-speed data access.

The initiative highlights a growing divide in the tech industry between the pursuit of increasingly sophisticated consumer products and the practical application of AI for global development. While the World Bank emphasizes that these smaller models are essential for progress, it notes that the success of this strategy depends on the ability of these nations to adapt the technology to their specific local contexts. The report suggests that focusing on these smaller, more manageable systems could provide a more sustainable path for economic growth in regions that currently lack the infrastructure to support the massive AI systems dominating the global market.