India faces one of the world’s most persistent food security challenges. About 9.8% of people in India experienced undernourishment in 2023-2025, while rising food costs continue to push a healthy diet further out of reach for millions of families.

India has fought this crisis since the early 2000s, and the effort shows results: undernourishment fell from 21.1% in 2004-2006 to 9.8% in 2023-2025. Affordability tells a harder story: 520.1 million people, 35.5% of the country, could not afford a healthy diet in 2025, and that diet now costs nearly 48% more than it did in 2017. Better crops, fewer losses and stronger food programs will all be needed to keep that progress going.

Farming and Undernourishment in India

More than half of India’s farmland still depends entirely on rainfall, so a single erratic monsoon, heatwave or drought can devastate a season’s harvest. Decades of policy have rewarded rice and wheat above all else, and that focus has pushed pulses, oilseeds and nutrient-rich millets to the margins while draining groundwater in the northwest. Farmers also lose substantial amounts of food after harvest: a 2022 government-commissioned study estimated losses of 3.89% to 5.92% for cereals, 6.02% to 15.05% for fruits and 4.87% to 11.61% for vegetables. Programs like NFSA, PM-POSHAN, ICDS and Rice Fortification secure calories and micronutrients well, but they still struggle to put protein, dairy and fresh greens on every table.

AI in agriculture is emerging as one answer to these gaps. The World Bank Group’s AgriConnect initiative, which aims to transform farming for 300 million smallholders worldwide, equips farmers with digital tools that work even on basic smartphones and in areas with weak connectivity, and those tools deliver fast, practical advice for managing crops day to day. Farmers who use them can head off pest outbreaks before they spread, time their harvests more precisely and reach government aid programs more easily.

Kerala’s KATHIR Platform

Kerala’s state government built KATHIR together with the World Bank Group, and the platform already holds data on more than 3 million farmers while mapping over 1.1 million hectares of cropland.

The platform packs in a wide range of tools: a 10-day weather forecast, sowing and irrigation advice, fertilizer recommendations, and early alerts for pests, crop stress and extreme weather.

KATHIR operates on two levels at once. Farmers get quick, usable guidance for their own fields, while the state government gains satellite crop tracking, harvest mapping, farmer identification and land enumeration, tools that let officials see exactly where support is needed and act on it.

Maharashtra’s MahaVISTAAR-AI

Maharashtra took a different route with MahaVISTAAR-AI, a generative-AI chatbot that answers farmers through voice or simple text, no typing skills or app literacy required. Farmers have downloaded the tool more than 3 million times, and the Korea World Bank Partnership Facility helped fund its development.

Looking Ahead

The World Bank Group plans to push these platforms well past crop advice. Future versions will help farmers complete official paperwork and connect them directly with banks and other institutions willing to design financial products built around their needs, cutting through bureaucracy that keeps many smallholders locked out of credit.

India’s experiment sits inside a much larger regional push. Google DeepMind built its Agricultural Landscape Understanding (ALU) and Agricultural Monitoring and Event Detection (AMED) models to map field boundaries and track farm activity, and those models played an early role in shaping India’s approach. DeepMind now extends that same support to countries across Asia-Pacific and Africa, work that matters given estimates that global food production must grow by as much as 50% by 2050 to feed 9.7 billion people.

AI in agriculture gives farmers who have long gone without expert support a direct, low-cost line to exactly that. If platforms like KATHIR and MahaVISTAAR-AI keep growing, and if banks and governments build the financial tools the World Bank Group has planned, AI could become an important tool for supporting rural livelihoods and reducing poverty India.

– Juan Guillermo Rueda Castrillón

Juan is based in Barcelona, Spain and focuses on Technology and Solutions for The Borgen Project.

Photo: Flickr