Making Artificial Intelligence work where it matters most

As leaders convene in New Delhi for the AI Impact Summit next week, Artificial Intelligence (AI) is entering a decisive phase.

- Johannes Zutt Sangbu Kim

Rapid advancements in frontier Al 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 Al accessible and practical now, so it can expand markets and power jobs for those who are often left behind.For developing countries, small Al 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 cuttingedge 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 Al is revolutionising 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.