The pandemic’s financial impact is drastically lowering expectations for economic growth in India. The road to prosperity will require a digital transformation, especially in the country’s agricultural sector. 70 percent of rural households rely on farming for their livelihoods, according to the U.N. Food and Agriculture Organization. Looking to help those farmers Indian nonprofit Wadhwani AI has partnered with Google’s philanthropic arm. The first project: redesigning the app CottonAce to take on pests such as the pink bollworm.
Cotton is a big deal in India, with 6 million local farmers growing the plant, but the cash crop is particularly vulnerable to the pink bollworm. This pest is estimated to wipe out up to a third of the country’s cotton crops each year. The insects are hard to detect visually, but Wadhwani AI aims to detect them before they enter the boll (the seed-bearing part of the plant)—which could be a game-changer for the community.
“Small farmers worldwide depend heavily on government and nonprofit programs to figure out what to do at every step of the crop cycle,” said Raghu Dharmaraju, vice president, products and programs, Wadhwani AI, Mumbai, India. “By using AI to augment human capabilities and overcome systemic challenges, we can help millions of farmers. Pest management is just the beginning.”
As recipients of a US$2 million grant, the Wadhwani AI team spent six months with nine of the tech giant’s uber experts known as Google Fellows to relaunch and redesign the app CottonAce. The redesign included developing more in-depth UX research and building a new infrastructure to speed up the AI learning model. The AI model was trained on 30,000 images and sought to both identify the presence of the pink bollworm and predict its trajectory. By fine tuning the algorithm agricultural experts are now able to deliver actionable advice on how to best eradicate the insects without the widespread use of pesticides.
To gather the data required for the app, farmers first take a photo of a pest trap containing the bugs. The app—which can work without an internet connection—then verifies the image and classifies and counts the pests. Based on the data, the AI model can predict where the next generation of eggs and larvae will settle before they become an infestation. The information is then sent to the agricultural experts, who recommend how to best counter the bugs.
Many of these experts had already been providing assistance to village farmers but were often unable to send back suggestions in time to save the blighted crops. When it came time to relaunch the app, the project team leveraged that relationship to build trust and secure buy-in for the new technology.
Throughout India’s harvest last June to September, CottonAce pilot programs were deployed across three states. An independent assessment found that small farms who used the app and followed the recommended advice saw jumps in profit margins of up to 26.5 percent and a drop in pesticide costs of up to 38 percent. The nonprofit has plans to deploy the app in full during this year’s harvest, aiming to deliver advice to 300,000 farmers this year—and 2 million farmers total by 2022.