A Global Pivot to 'Small AI'
In a significant move, the World Bank Group has launched its first comprehensive digital strategy, urging a focus on what it terms 'small AI'. This directive, announced as part of a broader initiative to expand digital access and create jobs, signals
a major rethink of how artificial intelligence can best serve developing economies. Instead of chasing the massive, resource-intensive 'giant models' that dominate headlines, the bank is advocating for practical, nimble, and affordable AI applications tailored to local needs. World Bank Group President Ajay Banga described it as the promise of 'AI for the people', emphasizing tools that can help countries make scarce resources go further.
The Problem with Giant Models
While large language models (LLMs) and other frontier AI systems are incredibly powerful, they present significant barriers for developing nations. Training and running these models requires immense computational power, specialized talent, and vast amounts of electricity, resources that are often scarce. Africa, for instance, accounts for less than 1% of global data-centre infrastructure. This concentration of power in a few advanced economies risks deepening the digital divide, leaving many countries dependent on technology that isn't built for their specific challenges. The high cost and infrastructure demands make building and deploying these giant models unrealistic for most emerging economies in the near term.
What 'Practical AI' Looks Like in Action
So, what does this 'small AI' look like? It’s about lightweight, task-specific applications that can run on everyday devices like smartphones, often with limited internet connectivity. These tools are designed to solve immediate, local problems by flourishing on smaller datasets and using minimal resources. Examples are already making a tangible impact. In Kenya, an app allows farmers to photograph a crop leaf and get an instant disease diagnosis without needing a constant internet connection. In Ghana, an AI-powered math tutor delivered via WhatsApp is producing significant learning gains for just a few dollars per student. In India, the focus is on providing farmers with critical soil analysis and market information in their local languages, a task made possible by accessible AI on a mobile phone.
Why This Matters for India and Other Nations
This shift is particularly relevant for India, where a vibrant startup ecosystem is already focused on solving local challenges. By prioritising practical AI, the World Bank’s strategy validates and encourages the development of solutions that cater to the unique needs of agriculture, healthcare, and education in emerging markets. It's a move away from simply adopting technology from the West towards adapting it to local contexts, languages, and infrastructure realities. For the 1.2 billion young people expected to enter the workforce in developing countries over the next decade, this approach to technology can help create jobs and improve productivity across key sectors. It empowers local entrepreneurs and small businesses to leverage AI for everything from inventory management to customer service without prohibitive costs.
Building the Foundations for the Future
The World Bank isn't just calling for this change; it's actively supporting it. Alongside the new strategy, it announced the 'AI Infrastructure and Readiness Initiative'. This program aims to help governments in Africa and Latin America access affordable, pay-as-you-go computing power, offering a practical alternative to building expensive national data centres. The overall strategy focuses on four pillars: expanding connectivity, making devices and data more affordable, strengthening digital skills, and preparing countries for AI. The message is clear: to truly benefit from the AI revolution, countries must first build the foundational 'analog complements'—reliable infrastructure, strong education systems, and robust data privacy frameworks.
