‘Farmers are gradually building trust in AI-based solutions’
The Artificial Intelligence (AI)-based pest detection solution for cotton developed by Wadhwani AI (formerly known as Wadhwani Institute for Artificial Intelligence) was recently integrated into the National Pest Surveillance System. Set up in 2018, Wadhwani AI has positioned itself “in the middle,” bridging the gap by combining research capabilities with market-focused applications, particularly in agriculture, education and health, says CEO, Shekar Sivasubramanian. “Our philosophy is grounded in understanding real-world societal problems and creating AI solutions quickly and effectively, often within three months. Promoting AI adoption in agriculture requires simplifying technology interfaces, integrating with existing systems, and consistent user training,” Sivasubramanian tells businessline in an interaction. Excerpts.
What are the technologies developed for the agri sector at Wadhwani AI? Have they been commercialised?
Wadhwani AI has developed several AI-driven solutions for agriculture, including Cotton Ace, a pest detection system. Cotton Ace uses computer vision to identify Pink Bollworm and American Bollworm infestations from photos of trap catches, which farmers collect using pheromone traps and empty onto white paper or cloth for analysis. After detecting the infestation level, the application provides farmers with tailored pesticide spray recommendations based on the severity of the infestation. It was the first reasonably large-scale, ground-level intervention for small-scale farmers. Cotton Ace has been integrated into the India government’s National Pest Surveillance System (NPSS). To address broader challenges, Wadhwani AI developed an integrated agriculture news monitoring system that uses AI to extract data from multilingual news reports on weather changes, pest infestations and water quality issues. This information is transformed into actionable insights and shared with government departments. Additionally, Wadhwani AI introduced an AI-powered grievance redressal chatbot for the PM Kisan grievance redressal system, streamlining the resolution of farmer complaints and benefiting millions of farmers. While these technologies are not commercialised for profit, they are scaled in partnership with government entities.
Besides cotton and rice, what are the specific crops that Wadhwani AI is targeting?
Besides cotton and rice, Wadhwani AI is working with the government to address other crop-pest combinations. We have done the same thing for rice. And for some other crop and pest combinations, the government says we have some data. While specific crops are not named, the focus is on tailoring AI solutions for different crops and pests based on government data and expertise. The institute emphasises scalability and integration of their pest management frameworks across multiple crops.
How is AI benefiting the Indian farmers?
Wadhwani AI’s initiatives provide Indian farmers with timely, data-driven advisories for improved agricultural practices. For instance, the Cotton Ace solution provides instant pest advisories. The moment a farmer takes the photograph, immediately they get the response. Farmers are gradually building trust in these systems. Most of the farmers told us they are very happy that we are bringing such technologies to them. Additionally, grievance redressal systems have reduced delays in government payouts for schemes like PM Kisan, ensuring farmers receive their entitlements faster.
How do you see the adoption of AI in Indian agriculture by the farmers in general?
AI adoption among Indian farmers is growing but faces challenges due to digital literacy and infrastructure gaps. These are people for whom you need to give them prescriptive solutions, minimal interfaces which is a different design paradigm compared to what you may face in urbanised cities. Building trust is a gradual process, as it takes time. Each of these initiatives takes some time. The importance of demonstrating tangible benefits is to encourage adoption. The unintended consequence was to build trust in the ecosystem into using it. They wanted to see what this is? What is this game? What is this whole thing all about? Farmer behaviour, such as sporadic visits to fields due to weather conditions, also affects consistent use of technology, requiring adaptations in implementation.
What needs to be done to promote the use of technologies like AI in agriculture?
Promoting AI adoption in agriculture requires simplifying technology interfaces, integrating with existing systems, and consistent user training. We have to understand and accept that the usage of technology also happens at a certain pace. This is not a set of people who are digitally literate. There is a need for iterative development. You need to know how to work with the marketplace. And we have design techniques, so that it will augment the performance of the system. Wadhwani AI also focuses on human-centred design and training. Partnerships with governments and organisations are vital to scale these technologies and address operational challenges.