As leaders convene in New Delhi for the AI Impact Summit next week, Artificial Intelligence (AI) is entering a decisive phase. Rapid advancements in frontier AI models — powered by large models, vast datasets and advanced computing capacity — are expanding the boundaries of what technology can do. At the same time, an urgent story is unfolding across much of the developing world that is focused on making AI accessible and practical now, so it can expand markets and power jobs for those who are often left behind.
For developing countries, small AI offers immediate and transformative value. These practical, affordable, and context-specific AI solutions are designed to solve specific problems — like crop pest detection, health care triage, or tutoring — without requiring cutting-edge digital infrastructure. These AI solutions typically rely on smaller models and data sets, run on everyday devices, can operate with intermittent connectivity, and are delivered through channels people already use.
Small AI is revolutionizing development by enabling governments, farmers, health workers, and small businesses to make better decisions faster — using tools that are accessible and affordable. In Kenya, the Nuru app enables farmers to capture images of infected crop leaves and receive immediate diagnostic feedback, even in low-connectivity settings.
Education platforms such as Bangladesh’s Shikkhok integrate AI into mobile applications that function offline and across multiple languages, extending learning to more students. These experiences reflect a growing global shift toward small AI adoption in developing countries. In India, small AI is already expanding access to services and opportunities at scale. India’s digital public infrastructure (DPI), built on digital identification (Aadhaar) and real-time payment system (UPI), has created a foundation where AI can be embedded into everyday systems at scale.
And AI that responds to voice is lowering transaction costs, reducing language and literacy barriers, and enabling AI-driven services to operate across multilingual, low-bandwidth, and informal markets at scale. In Andhra Pradesh and Telangana, the World Bank is supporting personalized learning tools by leveraging AI to provide children with individualized learning plans that adapt to each student’s pace. In Uttar Pradesh, AI-based drone technologies are helping fish farmers to scale their produce from local to global markets, boosting incomes and creating more jobs. These successes are powered by small AI designed for India’s context, built on strong digital public infrastructure, open systems, and aligned with public purpose. And AI-based drone technologies help fish farmers to scale their produce from local to global markets, boosting incomes and creating more jobs. These successes are powered by small AI designed for India’s context, built on strong digital public infrastructure, open systems, and aligned with public purpose.
Small AI, however, is not the end of the AI journey — it is the entry point. As countries build capability and demand through practical AI applications, they can progressively adopt more advanced AI systems and unlock deeper gains in productivity, innovation, and growth.
This progression must be supported by responsible governance from the outside. Strong safeguards for data transparency, accountability, and safety help build public trust and ensure AI systems are used in ways that are fair, reliable, and aligned with public interest.
Across the developing world, countries are keen to pilot AI solutions that deliver practical results, but they often need support to identify proven use cases, adapt solutions to local conditions, and scale them responsibly.
In India, the World Bank Group is supporting sandboxes to experiment with AI in safe environments where solutions can be tested. Around the world, the World Bank Group is helping countries learn from one another, sharing knowledge about what AI solutions actually deliver impact in developing countries so governments and innovators can learn faster and from evidence. AI is advancing at an extraordinary speed. But moments of rapid technological change also demand reflection. We are at a pivotal moment to make AI more accessible and locally grounded — unlocking broader opportunity and stronger development impact.
The next chapter of AI is defined not only by what technology can achieve, but by what it delivers for people, especially the millions who have too often been last to benefit from the world’s biggest breakthroughs.
Johannes Zutt is vice president for the South Asia Region and Sangbu Kim is vice president for Digital and AI of the World Bank. The views expressed are personal
