AI for agriculture Scaling Intelegence for food and climate resiliance
20 Feb 2026 10:00h - 11:00h
AI for agriculture Scaling Intelegence for food and climate resiliance
Session at a glance
Summary
This discussion focused on the integration of artificial intelligence in agriculture to enhance food security and climate resilience, featuring key stakeholders from the Indian government, World Bank, and research institutions. The session was part of the India AI Impact Summit, with Maharashtra’s Chief Minister Devendra Fadnavis presenting the state’s pioneering Maha Agri AI Policy 2025-2029 as a model for AI-driven agricultural transformation.
Dr. Devesh Chaturvedi from the Ministry of Agriculture outlined how AI systems like Bharatvistar are addressing the fragmentation of digital services by providing farmers with integrated access to weather advisories, crop guidance, market information, and government schemes through a single platform. He emphasized that AI serves as a “digital friend” to farmers, offering personalized advisories based on farmer IDs and digital crop surveys, with plans to expand to all Indian languages. The system aims to complement traditional extension services by reaching farmers at unprecedented scale and speed.
Johannes Zutt from the World Bank highlighted the revolutionary potential of AI for farmers while emphasizing the need for proper governance, interoperability, and accessibility. He stressed the importance of government foundations in ensuring credible, science-backed advisories and the private sector’s role in developing creative applications. Dr. Soumya Swaminathan raised critical concerns about inclusivity, particularly for women farmers who often lack land ownership documentation and may be excluded from AI systems built on existing datasets. She advocated for reducing drudgery for women farmers and maintaining human oversight in AI systems.
Shankar Maruwada from EkStep Foundation drew parallels between current AI adoption and historical agricultural revolutions, emphasizing the importance of open, interoperable systems based on Digital Public Infrastructure principles. The discussion concluded with plans for the AI for Agri 2026 Global Conference in Mumbai, positioning India as a leader in responsible AI deployment for agriculture at population scale.
Keypoints
Major Discussion Points:
– AI Integration in Agriculture Systems: The discussion focused on implementing AI-powered platforms like Mahavistar and Bharatvistar to provide farmers with personalized advisories, weather information, pest alerts, and market intelligence in local languages, moving beyond pilot projects to population-scale deployment.
– Digital Public Infrastructure (DPI) for Agriculture: Emphasis on building interoperable, open-source agricultural systems including farmer IDs, digital crop surveys, and data exchange platforms (like MahaEGX) that can integrate diverse datasets while maintaining data governance and farmer consent.
– Inclusion and Gender Equity in AI Agriculture: Strong focus on ensuring women farmers are not left behind in the AI revolution, addressing challenges like land ownership documentation, reducing drudgery, and designing systems that work for smallholders and tribal communities.
– Center-State Collaboration Framework: Discussion of how national and state governments can work together to scale AI solutions while maintaining interoperability, with Maharashtra’s Maha Agri AI Policy 2025-2029 serving as a model for other states.
– Global South Leadership and Knowledge Sharing: Positioning India as a leader in responsible AI deployment for developing countries, with platforms like the upcoming AI for Agri 2026 conference facilitating south-south knowledge exchange and collaboration.
Overall Purpose:
The discussion aimed to chart a roadmap for transitioning from pilot AI projects to population-scale implementation in agriculture, focusing on building trustworthy, inclusive, and interoperable AI systems that can serve India’s 500+ million farmers while establishing India as a global leader in agricultural AI for developing nations.
Overall Tone:
The tone was consistently optimistic and visionary throughout, with speakers expressing confidence in India’s ability to lead agricultural AI transformation. The discussion maintained a collaborative spirit, emphasizing partnership between government, private sector, academia, and international organizations. While acknowledging challenges like digital literacy and inclusion, the overall sentiment remained forward-looking and solution-oriented, with concrete examples of successful implementations reinforcing the positive outlook.
Speakers
Speakers from the provided list:
– Vikas Chandra Rastogi: Secretary of Ministry of Agriculture and Farmers Welfare, Government of Maharashtra – leads the session and panel discussion
– Devendra Fadnavis: Honourable Chief Minister of Maharashtra – provides keynote address on AI in agriculture and Maharashtra’s Maha Agri AI Policy 2025-2029
– Johannes Zutt: Regional Vice President, World Bank – brings global perspective on development finance and AI applications in agriculture
– Dr. Devesh Chaturvedi: Secretary, Ministry of Agriculture and Farmer Welfare, Government of India – leads national efforts in agriculture and digital agriculture mission
– Dr. Soumya Swaminathan: Chairperson of Dr. M.S. Swaminathan Research Foundation – global leader in science, champion for sustainable research and advocate for women farmers in agriculture
– Shankar Maruwada: Co-Founder and CEO of Ekstep Foundation – pioneer in building digital public infrastructure, co-developer of Mahavistar platform
Additional speakers:
– Mr. Ramesh Chaturvedi: Secretary of Ministry of Agriculture and Farmers Welfare (mentioned in opening but appears to be the same person as Dr. Devesh Chaturvedi)
– Mr. David Rupadnavi: Mentioned as Chief Minister but appears to be an error in transcription
– Shri Ashish Elarji: Honourable Minister (role/expertise not specified)
– Shri Nitesh Raneji: Minister (role/expertise not specified)
Full session report
This comprehensive discussion at the India AI Impact Summit brought together key stakeholders to chart a roadmap for implementing artificial intelligence in agriculture at population scale while ensuring inclusivity and responsible governance. The session featured Maharashtra’s Chief Minister Devendra Fadnavis presenting the state’s Maha Agri AI Policy 2025-2029 as a blueprint for AI-driven agricultural transformation, alongside insights from Dr. Devesh Chaturvedi, Secretary of the Ministry of Agriculture and Farmer Welfare, Dr. Soumya Swaminathan, and other distinguished panelists.
Strategic Vision and Four-Pillar Framework
Chief Minister Fadnavis articulated a compelling vision positioning AI as a carefully governed tool requiring trusted data and ethical frameworks. His emphasis that “AI is not magic” and “without trust, scale will not happen” established the foundational principle that technological advancement must be grounded in robust governance structures. He outlined four essential pillars for AI implementation: transparent, auditable, and explainable AI systems; open and interoperable digital infrastructure; strategic investment and scaling mechanisms; and inclusion with gender equity.
Maharashtra’s strategic approach represents a shift from pilot projects to population-scale implementation. The state’s Mahavistar platform has reached over 2.5 million farmers with multilingual personalized advisories, market intelligence, pest alerts, and government services access. Significantly, the platform recently integrated Bhili, the first tribal language in the country, demonstrating commitment to linguistic inclusion. As Fadnavis noted, this success proves that “farmers are ready for AI when AI is designed for them.”
The Chief Minister positioned Maharashtra as a “laboratory” for India’s AI ambitions, leveraging the state’s diverse agricultural conditions, research institutions, and startup ecosystem to develop solutions applicable across India and the Global South.
National Digital Infrastructure and Three-Phase Evolution
Dr. Devesh Chaturvedi provided crucial insights into the national framework supporting AI integration, candidly acknowledging the challenge of “digital red tapism” where multiple disconnected systems create complexity rather than solutions. The development of nearly 9 crore farmer IDs represents foundational progress in creating digital public infrastructure, enabling personalized AI advisories through platforms like Bharatvistar without requiring farmers to repeatedly verify credentials.
Chaturvedi outlined a three-phase evolution for agricultural AI: first, integrated advisory services consolidating weather, crop guidance, market information, and government schemes; second, highly tailored advisories based on comprehensive farmer data integration; and third, predictive models using historical data for weather patterns, market conditions, and farming decisions. He highlighted successful collaboration with Google’s CEO on predictive models serving 3.8 crore farmers using 100 years of meteorological data.
Crucially, Chaturvedi emphasized that AI serves as “additional to” rather than replacement for human extension services, with the AgriStack helping farmers access various government schemes and services seamlessly.
Global Development Perspective and Collaborative Framework
Johannes Zutt from the World Bank brought international perspective, drawing from his farming background to emphasize AI’s revolutionary potential for agricultural information access. He delineated clear roles: governments providing AI governance, interoperability standards, accessibility infrastructure, and digital literacy programs; private sector driving creativity and innovation in application development; and development institutions contributing financing and advisory services for quality validation.
India’s unique position for global leadership stems from its scale, diversity, and proven digital innovation. The country’s experience with varied languages, regions, crops, and farming conditions creates a natural laboratory for developing AI solutions applicable across developing countries, positioning India centrally in South-South knowledge exchange.
Inclusion and Gender Equity Imperatives
Dr. Soumya Swaminathan raised critical concerns about ensuring AI transformation strengthens rather than marginalizes vulnerable populations, particularly women farmers. She highlighted agriculture’s increasing feminization as men migrate to cities while women assume greater farming responsibilities, yet most land documents don’t bear women’s names, risking systematic exclusion from AI systems built on official data.
Swaminathan advocated for “humans in the loop” approaches rather than complete automation, referencing a film featuring a tribal woman from Jharkhand that illustrates this principle. She called for AI systems designed with women farmers through co-creation rather than mere consultation, citing examples like the Fisher Friendly Mobile App and Women Connect app from the MS Swaminathan Research Foundation.
She proposed institutional safeguards including farmer representation in AI evaluation committees, ongoing research and feedback loops, and rigorous evaluation methodologies similar to clinical trials for validating AI applications’ effectiveness and farmer benefit.
Technical Architecture and Scaling Through Digital Public Infrastructure
Shankar Maruwada from EkStep Foundation provided the technical framework for scaling AI solutions through digital public infrastructure principles. Using a railway analogy, he explained how open protocols and interoperable networks enable diverse stakeholders to contribute data, technology, policy, and research while maintaining system coherence, similar to how unified rail gauge enables cargo movement across India.
The design principles for inclusive AI systems prioritize accessibility for marginalized users, with specifications that illiterate farmers using basic feature phones should access services through voice interaction in local dialects. Mahavistar’s nine-month development involved collaboration across multiple institutions including the Government of Maharashtra, India AI Mission, Bhashini, IIT Madras, IIIT Hyderabad, World Bank, and Google.
Unlike traditional technology requiring perfection before deployment, AI systems improve through use, creating virtuous cycles of better models, enhanced data, and increased usage. This iterative approach enables minimum viable product launches followed by continuous enhancement based on user feedback.
Implementation Challenges and Adaptive Solutions
The discussion revealed sophisticated understanding of AI deployment challenges. The integration of diverse agricultural data demonstrates practical approaches to interoperability while maintaining data governance and farmer privacy. This enables researchers, institutions, startups, and government departments to access relevant datasets for innovation while protecting individual farmer information.
The expansion of Bharatvistar to all Bhashini-supported languages within three to six months demonstrates commitment to linguistic inclusion at national scale. The platform’s integration of weather advisories, crop guidance, market information, and government schemes into a single accessible interface exemplifies the “DPI is the new UPI” principle for agriculture.
Global Leadership and Knowledge Exchange
India’s positioning as a leader in responsible AI deployment for developing countries emerged as a central theme. The upcoming AI for Agri conference in Mumbai on February 22-23 represents institutionalization of this leadership role, providing platforms for South-South knowledge exchange and collaborative innovation. A global compendium of AI applications in agriculture, released in partnership with India AI Mission, World Bank, and Wadhwani AI, documents successful deployments across Africa, Asia, and Latin America.
Maharashtra’s invitation to venture capital funds, impact investors, multilateral development banks, and philanthropic foundations reflects understanding that scaling AI solutions requires diverse financial partnerships. The state offers compelling value propositions including diverse agro-climatic conditions, leading agricultural universities, AI research centers, vibrant startup ecosystems, and clear regulatory frameworks.
Future Commitments and Institutional Partnerships
The discussion concluded with concrete commitments for AI democratization. EkStep Foundation declared intentions to create 100 AI diffusion pathways by 2030 across different sectors, countries, and continents. The development of traceability digital public infrastructure as a replicable model addresses growing demands for supply chain transparency, food safety, and export competitiveness.
Collaboration between Maharashtra government and the MS Swaminathan Research Foundation on women’s rights in farming and nutritional security initiatives provides institutional mechanisms for addressing inclusion challenges. The recognition that AI must serve as a “digital friend” to farmers—continuously available, linguistically accessible, and contextually relevant—provides clear design principles for future development.
Strategic Implications and Path Forward
This discussion represents mature understanding of AI implementation challenges and opportunities in agriculture. The convergence of technical capability, policy frameworks, institutional partnerships, and inclusion principles creates favorable conditions for responsible AI scaling. The emphasis on trust, governance, and accountability alongside technological innovation reflects lessons from India’s digital transformation journey.
The strong consensus among diverse stakeholders on fundamental principles, while maintaining healthy debate on implementation approaches, suggests robust foundations for the ambitious transformation envisioned. The upcoming AI for Agri conference will provide crucial opportunities to translate these strategic discussions into operational partnerships and concrete implementations benefiting millions of farmers across the developing world.
The balance between rapid deployment and careful evaluation, between automation and human involvement, and between innovation and inclusion represents sophisticated policy thinking that positions India to influence global approaches to agricultural AI while ensuring that technological advancement serves the most vulnerable farming communities.
Session transcript
Mr. Ramesh Chaturvedi, Secretary of Ministry of Agriculture and Farmers Welfare. Sir, please come onto the stage. Our Honourable Chief Minister, Mr. David Rupadnavi is here. Good morning, sir, and welcome. May I also invite Mr. Johannes Jutt, Regional Vice President, World Bank, onto the stage, please. Honourable Chief Minister of Maharashtra, Mr. Devendra Fadnavis, Honourable Minister. Shri Ashish Elarji, Shri Nitesh Raneji, our distinguished guests from India and around the world. Very good morning. On behalf of the government of Maharashtra, I welcome you to the session on Using AI for Food and Climate Resilience. Agriculture is at a turning point. Climate change is making farming riskier, resources are limited and markets are changing quickly. However, there is an opportunity.
Digital tools and AI are advancing fast. Our goal is not just to use AI tools. We must build intelligence into our public systems to help everyone. For India, the change is essential. It is the key to food and nutrition security, higher farmer incomes and a stable economy. India is a country with a strong economy. India has shown that digital systems work when they are open and well -governed. Our next step is to bring AI into this framework in a responsible way. Under the leadership of Honorable Chief Minister of Maharashtra, the state has launched the Maha Agri AI Policy 2025 -2029. This policy uses AI for pharma advisory services, market information, data exchange, product traceability, innovation and research, and creating capacities of stakeholders.
We are moving beyond pilots to projects at full scale. Mahavistar is the country’s first AI -powered network and information and advisory services. Today, Mahavistar is being used by more than 2 .5 million farmers to get advisories in Marathi language, and recently the first tribal language in the country, Bili, has also been integrated into Mahavistar. AgriStrike is helping to bring AI into the market. It is helping farmers to get seamless access to various schemes and services. the Maha AgEx which is an open federated and consent driven architecture for data exchange it is helping us to bring diverse data sets together to get us a big picture. Agriculture is now a key part of India AI mission.
We are proud to work with the government of India to lead this change. I want to thank the Ministry of Electronics and Information Technology, Ministry of Agriculture Extra Foundation, the World Bank, MS Swaminathan Research Foundation, the Gates Foundation and all our partners for their support. It is now my duty to invite our Honourable Chief Minister to the stage. He will share his vision for using AI to strengthen our food systems and protect our climate. After the address of Honourable Chief Minister, we have a panel discussion with our distinguished panelists. Welcome.
A very good morning to all of you. Shri Devesh Chaturvedi ji, Rajesh Agarwal ji, Vikas Rastogi ji. Mr. Jonas Jett, Srimati Swaminathan, Shushankar Maruwada, my colleagues, Shriashi Shailar ji, Nitesh Rane ji. All the dignitaries present here, namaskar and good morning to everyone. It is my privilege to address this distinguished gathering at the India AI Impact Summit. And this important session. On AI in agriculture. We meet at a very defining moment across the world. Food systems are under strain. Climate volatility is intensifying. Water tables are falling. Soil health is deteriorating. Supply chains are fragile. And global markets are unpredictable. For countries from the global south, agriculture is not merely an economic sector. It is livelihood, social stability and national security.
India understands this very deeply. And under the visionary leadership of our Honorable Prime Minister Narendra Modi, India has placed digital public infrastructure and responsible infrastructure at the center stage of national development. The India AI mission is about using technology to deliver inclusion, transparency and scale Today, agriculture must sit at the heart of this mission Over half a billion Indians depend directly or indirectly on agriculture Yet, smallholders face fragmented information, rising input costs, climate uncertainty and limited access to credit and markets Traditional extension systems, however committed, cannot match the scale and the speed required Artificial intelligence changes this equation AI can provide hyperlocalization It can be used to predict and predict the future of agriculture It can be used to predict and predict the future of agriculture It can be used to predict and predict the future of agriculture It can be used to predict and predict the future of agriculture It can be used to predict and predict the future of agriculture It can be used to predict and predict the future of agriculture It can be used to predict and predict the future of agriculture It can be used to predict and predict the future of agriculture It can be used to predict and predict the future of agriculture credit scoring based on crop intelligence, transparent traceable supply chains, real -time market advisories.
But let me emphasize, AI is not a magic. As Honorable PM said in his inaugural session, AI must be built on trusted data, ethical governance, and public accountability. Without trust, scale will not happen. Last year, Maharashtra made a very clear and decisive strategic decision. AI in agriculture must not remain confined to demonstrations or pilots. It must reach millions. Under our Maha Agri AI policy 2025 -2029, we adopted Maha Agri AI policy 2025 -2029, we adopted a policy -led, ecosystem -driven model. built on openness and interoperability. Allow me to share what this has meant in practice. As rightly told by our Secretary Mahavistar, our AI -powered mobile platform delivers multilingual personalized advisories, market intelligence, pest alerts, and access to government services more than 2 .5 million downloads, acting as a digital friend to all these farmers.
This demonstrates one thing very clearly. Farmers are ready for AI when AI is designed for them. AI -based pest surveillance, crop sap integration is our mantra. By integrating, geospatial analytics, With post -surveillance, we have delivered early warnings to cotton -growing farmers, reducing crop vulnerability and finance risk. This is predictive governance in action. Agriculture data exchange is also one thing which is defining this step. We are building a statewide interoperable agriculture data exchange based on open standards and strong data governance. Data must empower farmers, not exploit them. Traceability digital public infrastructure in today’s global markets, the transparency is a mantra. We are unveiling a blueprint. For more information, visit www .fema .gov. a traceability DPI that will ensure end -to -end visibility across value chains enhancing food safety, export competitiveness and consumer trust and this is not proprietary.
It is being designed as a replicable public infrastructure model for India and the entire global south. In partnership with India AI Mission the government of Maharashtra, the World Bank and the Wadhwani AI, we launched a global call for AI use cases in agriculture. The resulting compendium of real world AI applications in agriculture was released in Delhi on 17th February 2026. This compendium documents successful AI deployments from Africa, Asia, Latin America and beyond. India is convening global knowledge for the benefit of the global south. As we move towards AI for Agri -2026 in Mumbai, our vision rests on four pillars. AI must be transparent, auditable and explainable. Open and interoperable digital infrastructure. Innovation cannot scale in silos.
Investment and scaling. Technology without capital remains just a theory. And inclusion and gender equity is also a mantra. Agri -2026. Is the international year of women in agriculture. AI solutions must be designed. with women farmers, not merely for them. Maharashtra today presents one of the most compelling agri -innovation ecosystems globally. 150 lakh hectares of cultivated land, diverse agro -climatic conditions, leading agriculture universities and AI research centres, a vibrant start -up ecosystem, a clear regulatory framework, and single -window facilitation for investors. We invite venture capital funds, impact investors, multilateral development banks, corporate innovation arms, and philanthropic foundations to partner with us. And in this partnership, we initiate a global partnership between Maharashtra and the United States to develop and leverage the technology to create a future for all.
Maharashtra is a partner of the International Development Fund. Maharashtra is a partner of the International Development Fund. Maharashtra is a partner of the International Development Fund. Maharashtra is a partner of the International Development Fund. Maharashtra is a partner of the International Development Fund. Maharashtra is a partner of the International Development Fund. Maharashtra is a partner of the International Development Fund. co -developing traceability DPI modules, investing in agri -tech startups, supporting digital literacy, especially among women farmers, building capacity in the rural AI ecosystems. When you invest in Maharashtra, you invest in scalable solutions for engaging economies worldwide, food security, climate resilience and AI governance are deeply connected. Countries that master AI -enabled agriculture will secure farmer incomes and strategic stability.
India has the scale, DPI and democratic governance model to demonstrate how AI can be deployed responsibly at population scale. Maharashtra is proud to be laboratory of that ambition. Friends, this satellite session is a declaration. We will move from pilots to platforms, from fragmented data to interoperable systems, from experimentation to execution, from intention to investment. The government of Maharashtra stands ready to collaborate with the government of India, with states, with global institutions, investors, researchers and farmer organizations. Let us ensure that AI becomes a force for
Thank you. Thank you, sir, for your visionary address. You always continue to inspire us to aim higher and achieve better. And under your leadership, I can assure you the Agriculture Department will rise to the challenge and serve the aspirations of more than 15 million farmers of the state of Maharashtra. Thank you so much, sir. We will now start the panel discussion in a few moments. Thank you. Thank you. Thank you. Thank you. Once again. Dr. Devesh Chaturvedi, he is the Secretary, Ministry of Agriculture and Farmer Welfare Dr. Chaturvedi leads our national effort in agriculture and farmers welfare Mr. Johannes Jett, he is the Regional Vice President, World Bank Mr. Jett brings a vital global perspective on development and finance from the World Bank Ms. Soumya Swaminathan, she is the Chairperson of Dr. M. S. Swaminathan Research Foundation Dr. Swaminathan is a global leader in science, a champion for sustainable research and a strong advocate for mainstreaming women farmers’ role in agriculture Mr. Shankar Maruwala, he is the Co -Founder and CEO of Agestep Foundation He is a pioneer in building digital public infrastructure that empowers women farmers to develop their own agriculture and empowers people at scale and I am very proud to say that the Government of Maharashtra and Agestep Foundation together have brought out Mahavistar, which more than 2 .5 million farmers are using today to get the advisories and information that they need on a daily basis.
The objective of this panel discussion is to move from vision to implementation. Specifically, we will deliberate on how to institutionalize AI within agriculture systems at scale, how to ensure inclusion, especially of women farmers and smallholders, how to build interoperable, trustworthy and sustainable AI governance ecosystems, and how to strengthen collaboration between the center, states, global institutions, industry and academia. The session is also an important precursor to AI for Agri 2026 Global Conference, where we will continue these deliberations in greater operational depth with governments, investors, investors. innovators and development partners. AI for Agri conference is being held in Mumbai on 22nd and 23rd of February at Jio World Convention Centre. With this context, let’s begin our discussion.
My first question is to Dr. Devesh Chaturvedi. Sir, under your leadership, the ministry has taken significant steps in advancing the digital agriculture mission and operationalizing the Agri -STEC framework. You are laying a strong digital foundation for the sector. As we now look at integrating AI more systematically into agriculture, how do you envision the central state collaboration framework, specifically to ensure that AI deployments are aligned with national architecture while allowing states the flexibility to innovate based on local agro -climatic and socio -economic context? And finally, how can we institutionalize this collaboration? to achieve population scale impact while mentoring interoperability and data trust. Thank you.
A lot of questions in the same question. So what I’ll do is I’ll just first take you through the initiatives. First of all, we deeply appreciate the leadership taken by Maharashtra under obviously the leadership of a vulnerable chief minister and with the agriculture department. They have done exceptional work in digital agriculture mission by developing farmer IDs and digital crop survey. And also they launched Mahavistar as a precursor of Bharatvistar. And recently on 17th, government of India have also launched one of the first integrated AI -based system for the farmers, which is Bharatvistar, which presently is undertaking, providing services, which is work through the app. Android based app as well as through mobile telephony on weather advisories ICR based crop advisories, pest advisories market information regarding various agriculture produced, traded in the Mondays and lastly the government schemes of government of India.
Now why is this important, AI is important in agriculture? Like we did a lot of, we started with digitalization of services, different services we had DBT we had online systems of applying for various common person, applying to the common service centers or through the mobile apps but what was felt was that while we had initiated this process to ensure that the bureaucratic red tapism is removed, what we were moving towards was a sort of digital red tapism because within our ministry different schemes had different apps and they had different ways of selection and within the state also horticulture had a different database of farmers, agriculture had a different database, animal husbandry has a different database, crop insurance has a different database.
So basically a farmer who has to avail so many services was, we felt that he or she was getting lost in which app to use and which one to use. And sometimes it becomes more difficult to avail the services through online systems or to get advisories than to go to a person and say, okay, tell me how to do it. So the whole idea was that once we have this AI -based system, we have a same platform for different applications and different advisories at a click of the button or maybe just as a voice. So that is the whole idea of shifting towards AI -based solutions. So now what we have initially in the first phase in the artificial intelligence system, the Bharat Vistar or the Mahavistar of Maharashtra, is that the advisories, the crop advisories, the weather advisories, schemes information about how to apply and what is the status of that application, and also the Monday rates.
All these have been put in the one platform. You can just make a – presently it is working in English and Hindi, but in the next three to six months we’ll be taking it towards all the Bhashani -related languages. And the next step is, as we mentioned, that the states are working together with us for the digital public infrastructure. So close to 9 crore farmer IDs have been developed. So what is a farmer ID? And you must have read the statement of Honourable Finance Minister, that DPI is the new UPI. So what is the basic – this agri -stack, which is a part of DPI, is that for agriculture is that we have – each farmer has a unique farmer ID with the back end of all the crops the person has sown, what is the land available to that person, all the data with the share of the land and the crops sown and the soil health card details if the soil health has been given.
So with these basic details available on the system, it empowers the farmer through that ID to avail services because it is already approved by the relevant authorities in the government. So the person does not have to or the authorities who are giving the services are not required to cross -verify the credentials of the farmer based on the record of rights or maybe the Girdhavari or whatever it was in the different states. So every state and Maharashtra is one of the leading states here. We are working together to have a saturation of farmer IDs and crop survey. And once this is there, then this AI will further transform into a very, very tailored advisory. So a person calls or gives the farmer ID or Aadhaar.
And at the back end, we will, based on the consent, access the details of where the farmer is from, what is the crop being grown, what is soil health conditions. And very targeted advisories will be given, which will be made operational in the next three to six months. So instead of pushing data which may not be of interest of the farmers, very specific, tailored, data for that farmer will be available based on integration. of digital public infrastructure with Bharat Vistar. And the third aspect will come when we do the predictive models. And we tried that and you must have remembered in the inaugural session when Google CEO mentioned about that predictive model which we did with about 3 .8 crore farmers.
We used 100 years data of IMD and a model to predict a monsoon for the next one month and for next week. And that prediction was fairly accurate and farmers we got the feedback to farmers did take the decision to sow and to irrigate based on the predictive model which was sent. And now we will expand the predictive models to ensure that we get more advisories of the market situation, of the weather situation which will help improving the decision making of the farmers and so that they can increase their productivity, reduce the cost. So that is the whole idea of AI in agriculture. And we hope that more and more farmers will adopt it and it will be a lot and it will be a lot exactly a replacement but a sort of additional to the human, we can say, extension services, which we find is not able to reach to the farmers because of the resource constraints of each state.
The extension machinery, the KVKs, all our state extension machineries, it’s very difficult to reach each and every farmer because of the fact that we can’t have a person sitting in each village reaching to each farmer. But AI, along with digital public infrastructure, along with the mobile and internet penetration in the various rural areas, will ensure that that gap is removed and we get more and more access to the farmers on services and advisors. That is the whole idea of having center and state interoperability. But I hope I have answered most of the questions which
As you rightly mentioned. AI systems are acting like a digital friend of the farmer so they are available at any point of time through multiple channels and in a language they understand in FEDSAR with ministries assistance we were able to get access to multiple images of pest and disease and with IIT Bombay we have been able to develop models where farmers can take a picture and they can find out what pest and disease is it and then ultimately what is to be done based on the knowledge created by agriculture universities and ICR institutions so I think there is a great opportunity for us the national government has the scale and the states have their own specific skill sets and knowledge together if they combine I think we can reach out to each and everybody in the farming sector.
Thank you sir I will move on to Mr. Johannes Jutt now the regional vice president of the World Bank the World Bank has been a long standing partner to both the government of India and the government of Maharashtra we have multiple projects going on concurrently as well as we have had in past as well. And these projects have been aimed at strengthening agriculture systems, climate resilience, and institutional capacity. As we move into the era where AI technologies are evolving at unprecedented speed, how can development partnership adapt to remain agile and responsive? In particular, how can we structure programs and technical assistance model that provide just -in -time support to central and state governments, enabling them to experiment, iterate, and scale AI solutions responsibly?
to be here today. So we’re on the cusp of a major revolution in how support to farmers and agriculture is happening. I actually grew up on a farm. I worked on a farm from the ages 10 to 21. I think every hour I wasn’t in school that I was actually at home. I was working in a farm. In some ways it feels paleolithic because we didn’t have computers. We had telephones that were connected to wires and our ability to get information about what was happening around us was extremely limited. We spent a lot of time trying to find out the things that today you can find out very, very quickly using small AI for agriculture.
And that’s truly revolutionarily empowering for farmers. But to make that work for farmers, there’s a lot of things that need to go right. And I think it’s worth reflecting a little bit about on the different roles that we have. Thank you. actors in the ecosystem have, starting obviously with government. My colleague mentioned a number of these things earlier. The government’s responsibility is principally on foundations, things like the governance of AI, the interoperability, accessibility, obviously ensuring that educational programs include appropriate types of skilling in the use of digital services. This is a big challenge in countries like India, where frankly there are still people who don’t have sufficient literacy to read what comes over a basic smartphone.
Ensuring that the research and extension that is provided through these small AI platforms, is credible, is trustworthy, is backed by science. I think that’s also extremely important. Of course, farmers will find out if they aren’t. but at high expense, right? So we want to make sure that they’re not being advised to do things that are negative for them. And then also looking at the cost of service, the connectivity, what does the farmer actually need to be able to link into these different types of platforms that give information? Because, of course, we’re often also talking about farmers who have very, very few assets and who may be essentially unable to stay permanently connected or even easily connected to the Internet.
They’re going to have very basic smartphones, et cetera. So the government has a lot of work to do in all of those areas. Then you can look at what can the private sector do. Now, one thing that the government needs to do is encourage, crowd in private sector capacity and capital. But once we turn to the private sector, what is the private sector’s principle? advantage. I think that there’s a lot of creativity in the private sector. So the actual applications that are being developed are being developed by individuals in the private sector with a passion for specific sorts of issues that are constraining farmer success. And, you know, that creativity will result in a number of different applications that will be aimed, in most cases, to help farmers overcome certain hurdles that they face.
And, you know, we can kind of let a thousand flowers bloom there and see what actually takes root. And it’s amazing what you start to see. Just yesterday, I was learning about an application in Morocco developed by a tomato farmer who was able to give advice about how much water tomato plants need simply by taking a picture. of the current tomato plant. Take a picture and it tells you how much water you actually need to give this plant, which obviously in a water -stressed environment is vital, vital information. And then, you know, there are roles for institutions like my own, the World Bank Group, which can help to provide some of the financing that helps develop these applications, and also the foundational backbone for artificial intelligence.
And we can also play a role at the advisory end, where we are helping to truth -test, if you like, the information that’s coming through different applications that are coming out of the AI sandbox in different contexts to make sure that it’s actually providing information that’s useful to the end beneficiary and enhancing from a productivity perspective at the farm level. Thanks.
I think you have rightly pointed out the role of innovation and research and what we see is we require high quality robust data to actually build upon that and as Honourable Chief Minister mentioned, MahaEGX is one step in that direction wherein we bring diverse data sets and make them accessible to researchers, academic institutions, departments and also start -ups and many of these start -ups we will see they are showcasing their innovations in AI for Agri conference in Mumbai. So we request all of you to please come there and see for themselves what kind of excitement they have and what kind of solutions are envisaged. I have one supplementary question to you. How do you see a platform such as AI Impact Summit as well as AI for Agri global conference contributing to deeper global collaboration and south -south knowledge exchange in this domain?
Thank you for that additional question. I mean, obviously, India is in a great position to lead the development of AI, particularly for developing countries where there are still significant challenges helping poor people to escape poverty permanently. India has demonstrated digital innovation for a long period of time already. It’s got an enormous population with a huge variety. The challenges of bringing farmer -appropriate data to the farmer’s fingertips in India are – I was going to say India is a microcosm of the rest of the world. It’s hardly a microcosm. It’s so huge. But because you have so many languages, so many different regions, so many different types of crops, and the starting conditions at the farm level are so incredibly varied, figuring out how to make AI at the farm level work, in India will automatically have a large number of spillover learnings for other countries around the world.
and because India after China and the United States is the country in the world that is best positioned actually to push all of this work forward and because it is itself a developing country, it’s very, very clear that it will have a central role to play in South -South learning for those reasons.
Thank you so much. I move on to Dr. Swaminathan. Dr. Swaminathan, your father, Professor M.S. Swaminathan, played a historic role in shaping India’s agriculture transformation during the Green Revolution, ensuring food security at a critical juncture in our history. Today, as we speak of a new phase of transformation driven by AI, we are again at an inflection point. You have consistently championed science -based policy, sustainability and the empowerment of women farmers. With 2026 being recognized internationally, as the year of women farmers, how can we ensure that AI -led agriculture transformation strengthens women’s agency? knowledge access and climate resilience and what institutional safeguards and design principles must be embed today so that this new technological revolution becomes equitable farmer centric and grounded in scientific integrity
Thank you very much for that question Vikasji not only is this year the international year of women farmer but we know that agriculture itself is increasingly being feminized with many men actually leaving farming to the women and migrating out to the cities for other opportunities so it is really essential to put women at the center of all that we are discussing and I think the chief minister today gave us a wonderful vision of what can be the future provided of course like you said that there are the guardrails there are the institutions there are the safeguards and the design principles that we think about from the very beginning so my father professor MS Parminathan used to say that the green revolution was not only about the seeds, of course the seeds played a very big role you know the high yielding varieties but it was about the entire ecosystem and the institutions that were developed at that time which included the outreach you know later on the Krishi Vigyan Kendras of course were developed but also the access to credit, the water, the fertilizers, the education, the empowerment and ultimately became a success because farmers realized the potential of it and took it on.
So what he used to say is that you know every technology, no technology is pro -poor or pro -rich or pro -woman or against women, it’s how we use that technology so it’s really like you said the inflection point today is how do we use this very powerful technology that’s come to us. So I think there are a few points here, you know, to make sure particularly that women farmers are not left behind. The first important fact is that women in India, the minority of them who have their name on the land document, so mostly it is in the man’s name, and Deveshji was telling me today that this is improving and that the latest census shows that perhaps at least a quarter of the properties are also in the name of women, either jointly or, but that still means that, you know, three -fourths of them don’t have.
And a system that operates basically on publicly available data will then leave out those whose data sets are not available. So I think it would be really important at the early stages itself to think about how women’s data can be incorporated because the algorithms are fed by the data we have. And so all of these advisories may be very suitable for a man who’s operating a tractor on a farm, but not at all relevant for a woman who’s still working with outdated instruments and trying to, you know, till her land. And particularly when we look at more remote areas, tribal areas, where women do a lot of the agriculture like millets, for example. Mostly it is women who grow millets.
And there’s still a lot of mechanization which is absent completely. It is all still very much done using traditional methods and tools, and it involves a lot of drudgery. So I would say that, you know, one of the benchmarks that I would look at is, is it reducing the drudgery and the workload on women farmers? Is AI helping to do that? So I think we also need to look at certain indicators for success. And you mentioned science. I mean, I’m a medical researcher, and the way that we evaluate products is by doing clinical trials, by examining the data and the evidence. And then recommending it for wider use. so again a note of caution would be to as we roll it out we need innovation certainly we also need to do the evaluation looking at inherent biases looking at who’s being excluded looking at are there unanticipated risks or side effects that we didn’t know about but most of all it’s this inclusion I think we don’t want those who are already left behind to be further left out so I think the ongoing research and data collection and feedback loops and most importantly having the voices of those for whom we are developing all these I think in the room I don’t think we have any farmers or women farmers so we are all discussing from what we know but if you are the farmer like you were saying working there and you know the constraints under which you are working so I think the women farmers and farmers in general must have a role they must be part of these committees that evaluate or make recommendations or make suggestions on improvement it has to be an iterative process I think any technology is as good as the application for which it’s developed I’ll give you one example of an app that the MS Farminathan Research Foundation developed for fisher women.
We had very successful app for fisher men called the Fisher Friendly Mobile App that won the UN Tech for Nature award last year. But fisher women were as usual left out and so the Women Connect app actually gives them on a tablet information that they need to sell because once the fishermen have come back from seeds, the women who have to do all of the post harvest and the same is true for crops or fruits or vegetables as well. So that connection to the market, of course information about pests and pathogens and when to buy what and what inputs to use but also being able to organize themselves. And I think women, there are many FPOs now and FPCs and SHGs made of women farmers, empowering them and giving them the knowledge and tools.
And the last thing I would say is we still need humans in the loop. I don’t think we should think that completely making everything run by machines is going to solve our problems. I think it’s risky there. And in a country like India, we also need employment. And so we should think of, and I don’t know how many of you have seen this film called Humans in the Loop. But it’s a tribal woman from Jharkhand who actually raises questions about the algorithm. It’s a very thought -provoking film. So I think Humans in the Loop is going to be important. We have our Krishis, Sakhis and so on. We need to empower them with these. So I think AI and all these digital tools, if they’re used in addition to the traditional knowledge and wisdom that people have and augment it and give them at the right time, at the right place, the knowledge they need, I think we can go a very long way.
Thank you.
Thank you, madam. You have rightly pointed out the need to be more sensitive and while developing systems for inclusivity. and to ensure that for whom they are being developed and they are in the loop and they are being consulted. In fact, the feedback mechanism that we have developed in Mahavista takes care of those requirements. I’m also very happy to share that Government of Maharashtra and Dr. M.S. Swaminathan Research Foundation are working together on some of these issues in terms of how to bring women’s right in farming at the center stage. How do we create bio -happiness using our universities and educational systems? And what kind of nutritional security we must look for? Because we have food security, but it’s the nutritional security that we must aspire for.
So we are happy to have support and assistance from MSSRF in that direction. My final question is to Mr. Shankar Maruwada. Mr. Shankar, ECSTEP has played a role. A foundational role in shaping India’s DPI landscape through open source platforms such as Sunbird. which has powered large -scale systems like Diksha, Mahavistar and Open Network Initiative built on backend protocol. These efforts have demonstrated how open standards and interoperable architecture can enable population -scale transformation that we are already seeing today. As we now enter the era of AI -driven public systems, how should we think about standardizing AI -based ecosystem in a similar spirit? How can we bring DPI into AI? And what architecture and governance principles are required to ensure interoperability, trust and sustainability in AI deployments across sectors such as agriculture?
Again, a whole lot of questions, but let me make my best attempt to answer those. More than 100 years ago, the world faced what was known as a Malthusian crisis, where Malthus, the economist, predicted that if we continue to grow, in the same way we’ll run out of land, we’ll run out of soil. We were a billion and a half then. We are 8 billion. Most of us may not even have heard of the Malthusian crisis. What happened? Someone called Haber and someone called Bosch created a miracle. Haber synthesized ammonia using high pressure and temperature and Bosch put it into an industrial process. That phenomena is now historically known as pulling bread out of air.
It took a lot of effort and as Soumya said, creation of a massive ecosystem. Germany, which pioneered this, lost that race to US. Because US did a better job of diffusing the technology safely to the farmers. They created the discipline of agriculture engineering. They created institutions like the Fertilizer Development Center. They helped. technology demonstrations to farmers to show them how synthetic ammonia could be used. By the way, 50 % of the nitrogen in our body comes from synthetic nitrate ammonia. That’s a fact. So we owe a lot to Haber and Bosch. China then took it on in 80s by buying 10 big plants from Kellogg, training 300 million farmers, showing them how to use synthetic fertilizers.
And they went on to be the global leaders in agriculture. India is at a point where if we learn the lessons from such past things, our green revolution, our DPI experience, we are at a pivotal point where the equivalent of pulling bread out of thin air is pulling intelligence from the earth and providing it to the farmer. this is again not science fiction Mahavistar, the pioneer along with Bharatvistar have taken the first steps to this so when a Mahavistar was designed to build off what Swami has said, it was designed with inclusion in mind inclusion diversity was not an afterthought because to solve for not just Maharashtra’s problems for India’s scale and diversity we need to think of the last person, the most discriminated in the remotest part of India and design systems that work for them we call that DPI now let me give you a specific example of this in Bharatvistar right from the beginning the design specs was we need an illiterate farmer to build off John’s point about digital literacy with a feature phone, not a smartphone, to be able to talk in his or her native language and native dialect.
Marathi itself has many dialects, right? Talk on the phone, like the way she is comfortable talking to another person. Ask the question, have a conversation, get a bunch of answers. That process took us the better part of nine months. Why? Because it’s not just AI. It’s data. It’s processes. It’s training the farm extension workers. It is having trust on will this work? What about the costing? Will I blow up my entire stage budget on a model, right? Do I have autonomy? Can I switch models out, in and out? These are very, very difficult questions. It took us in partnership with a whole lot of people. I mean, government of Maharashtra led the effort, but IndiAI mission, Bhashini, IIT Madras, IIIT Hyderabad, World Bank, Google many other providers everybody chipped in the little part of the solution now here is the best part because we all collaboratively invested in figuring out a solution there that solution could be deployed in Bharat Vistar with more confidence easily again the same challenges that secretary Chaturvedi talked about do we have the data he used a very nice phrase digital red tapism our data is in different formats what matters is the intent of the government of India which triggered the process which allowed Bharat Vistar to be launched the day before it’s a start data will get better, the systems will get better, usage will improve that will generate more data and then over time years the ecosystem will be built this we know from our experience what makes this happen what is that secret sauce the design principles it is the same as DPI what worked for DPI we are taking those same principles one open interoperable systems think networks and not just portals and platforms and siloed and fragmented systems what’s the best example of this the railways in India we have such a vast landscape but the rails are common every state can decide what it wants to move private public defense farming the Indian railways is just providing a backbone that allows everyone to do this there was a time when we had different rail gauges right now that sounds so silly but there was a time like that But India is showing that we don’t have to repeat those early mistakes in digital also.
By creating interoperable networks based on open protocols like Beacon, by collaborating with each other, one of us is bringing in data, somebody is bringing in technology, somebody is bringing in policy, somebody is bringing in research. These collaborative open networks and with the launch of Bharat Vistar puts India in a very unique and responsible position. Unique because we have these open rails, we have the experience of DPI. Responsible because it is a start. Unlike the technologies of the past where you perfect the technology and then deploy AI, you deploy something minimum to start and then evolution, models get better, data gets better, usage gets better. And then it gets better and better over time. that is the unique junction we are in in India what will that mean?
when I -CAR plugs into this network with its weather and pricing data that network makes it available to any state that wishes to turn on the supply from I -CAR when a private sector comes out with a very innovative app let’s say the tomato example that John talked about any state can say I like that I think I will have that made available to my farmers for the farmers they anyway trust the state they can go to the same app and now see this also there if the tomato app person wants they can go directly to each farmer very very expensive so Shared Rails allows us to spread innovation diffuse it very quickly through society keeping in mind both inclusion and inclusion and inclusion and rewarding innovation because innovation has to be rewarded.
And I want to end with a very simple analogy. When Edmund Hillary climbed Mount Everest, he made a lot of people believe it is possible. When Mahavistar was launched, it made the country believe that it is possible to make AI serve the farmer. And to that extent, the responsibility that Mahavistar, Maharashtra government and government of India has is to create these pathways for the rest of the country for the other states. At XTEP Foundation, Nandan Elekani, we made a declaration two days ago. We would like to see a world by 2030 where there are 100 such diffusion pathways, each created by a different set of people in different sectors, in different countries and continents, but each inspiring.
different AI pathways to safe impact at scale. And it’s a very exciting vision. It’s a very collaborative vision. If you all get together, we can also create miracles in our own lifetime. Thank you.
With that profound thought, we’ll conclude today’s panel discussion. I thank all the panelists. They have really opened a new vision in front of all of us. And we’ll invite all of you to AI for Agree conference in Mumbai on 22nd. Thank you so much. We don’t have question actually. Time for question. The next session is about to start.
Speech speed
106 words per minute
Speech length
1101 words
Speech time
619 seconds
AI must be built on trusted data, ethical governance and public accountability
Explanation
The minister emphasizes that artificial intelligence in agriculture should rest on reliable data sources, be governed by ethical standards, and be accountable to the public. This foundation is presented as essential for responsible AI deployment at scale.
Evidence
“As Honorable PM said in his inaugural session, AI must be built on trusted data, ethical governance, and public accountability.” [1]. “AI must be transparent, auditable and explainable.” [2].
Major discussion point
Vision and Policy Framework for AI in Agriculture
Topics
Artificial intelligence | Human rights and the ethical dimensions of the information society
Maha Agri AI Policy 2025‑2029 adopts an ecosystem‑driven model
Explanation
The policy adopts a government‑led, ecosystem‑driven approach to foster AI solutions for agriculture across Maharashtra. It aims to create an open, interoperable environment that encourages innovation and collaboration.
Evidence
“Under our Maha Agri AI policy 2025 -2029, we adopted Maha Agri AI policy 2025 -2029, we adopted a policy -led, ecosystem -driven model.” [16].
Major discussion point
Vision and Policy Framework for AI in Agriculture
Topics
The enabling environment for digital development | Artificial intelligence
Mahavistar delivers multilingual advisories to over 2.5 million farmers
Explanation
Mahavistar, an AI‑powered mobile platform, provides personalized, multilingual advisory services, market intelligence and pest alerts to more than 2.5 million farmers, acting as a digital companion for them.
Evidence
“As rightly told by our Secretary Mahavistar, our AI -powered mobile platform delivers multilingual personalized advisories, market intelligence, pest alerts, and access to government services more than 2 .5 million downloads, acting as a digital friend to all these farmers.” [18].
Major discussion point
Vision and Policy Framework for AI in Agriculture
Topics
Social and economic development | Information and communication technologies for development
Building a statewide interoperable agriculture data exchange on open standards
Explanation
The government is creating a state‑wide data exchange that follows open standards and strong governance, ensuring that data empowers farmers rather than being exploited. Open, interoperable infrastructure is portrayed as a cornerstone for trustworthy AI services.
Evidence
“We are building a statewide interoperable agriculture data exchange based on open standards and strong data governance.” [62]. “Data must empower farmers, not exploit them.” [64].
Major discussion point
Central‑State Collaboration, Data Interoperability and Digital Public Infrastructure
Topics
Data governance | Artificial intelligence
Scaling from pilots to platforms requires trust, investment and replicable public‑infrastructure
Explanation
The speaker stresses that moving beyond pilot projects to full‑scale platforms demands trust, investment, and a replicable public‑infrastructure model that can be shared with the global south. This transition is framed as essential for population‑level impact.
Evidence
“We will move from pilots to platforms, from fragmented data to interoperable systems, from experimentation to execution, from intention to investment.” [48]. “Without trust, scale will not happen.” [54]. “It is being designed as a replicable public infrastructure model for India and the entire global south.” [77].
Major discussion point
Vision and Policy Framework for AI in Agriculture
Topics
Artificial intelligence | The enabling environment for digital development
Invitation to venture capital, impact investors and multilateral development banks
Explanation
The minister calls on private capital and development finance institutions to partner in co‑developing traceability DPI modules, funding agri‑tech startups and supporting digital literacy, especially for women farmers.
Evidence
“We invite venture capital funds, impact investors, multilateral development banks, corporate innovation arms, and philanthropic foundations to partner with us.” [113].
Major discussion point
Role of Private Sector, Innovation, Financing and Global Partnerships
Topics
Financial mechanisms | The enabling environment for digital development
Global call for AI use cases and AI for Agri‑2026 conference
Explanation
A partnership with the India AI Mission, the World Bank and Wadhwani AI launched a global call for AI use cases in agriculture, and the AI for Agri‑2026 conference is positioned as a platform for South‑South knowledge exchange.
Evidence
“In partnership with India AI Mission the government of Maharashtra, the World Bank and the Wadhwani AI, we launched a global call for AI use cases in agriculture.” [29]. “India is convening global knowledge for the benefit of the global south.” [108].
Major discussion point
Role of Private Sector, Innovation, Financing and Global Partnerships
Topics
Artificial intelligence | International cooperation
Dr. Devesh Chaturvedi
Speech speed
163 words per minute
Speech length
1127 words
Speech time
414 seconds
Farmer IDs and the agri‑stack form the backbone of a unified data ecosystem
Explanation
The secretary explains that each farmer receives a unique ID linked to crop, land and soil‑health data, creating a foundational agri‑stack that eliminates fragmented databases and digital red‑tapism across schemes.
Evidence
“So what is the basic – this agri -stack, which is a part of DPI, is that for agriculture is that we have – each farmer has a unique farmer ID with the back end of all the crops the person has sown, what is the land available to that person, all the data with the share of the land and the crops sown and the soil health card details if the soil health has been given.” [34]. “what we had initiated this process to ensure that the bureaucratic red tapism is removed, what we were moving towards was a sort of digital red tapism because within our ministry different schemes had different apps…” [58].
Major discussion point
Central‑State Collaboration, Data Interoperability and Digital Public Infrastructure
Topics
Data governance | Capacity development
Integration of AI with digital public infrastructure and farmer IDs creates a single platform for services
Explanation
By combining AI, digital public infrastructure, and widespread mobile connectivity, the system can deliver weather, crop, pest, market and scheme information directly to farmers through a unified interface.
Evidence
“But AI, along with digital public infrastructure, along with the mobile and internet penetration in the various rural areas, will ensure that that gap is removed and we get more and more access to the farmers on services and advisors.” [30].
Major discussion point
Vision and Policy Framework for AI in Agriculture
Topics
Artificial intelligence | Information and communication technologies for development
Vikas Chandra Rastogi
Speech speed
110 words per minute
Speech length
1813 words
Speech time
985 seconds
Transition from pilots to full‑scale platforms is essential for population‑level impact
Explanation
The panelist stresses that moving beyond experimental pilots to fully operational platforms is necessary to achieve scale and deliver benefits to the entire farming population.
Evidence
“We are moving beyond pilots to projects at full scale.” [47]. “We will move from pilots to platforms, from fragmented data to interoperable systems, from experimentation to execution, from intention to investment.” [48].
Major discussion point
Vision and Policy Framework for AI in Agriculture
Topics
Artificial intelligence | Social and economic development
Maha AgEx provides a consent‑driven, open federated architecture for data exchange
Explanation
Rastogi describes Maha AgEx as an open, federated, consent‑driven data‑exchange architecture that aggregates diverse datasets, enabling researchers, startups and policymakers to access a comprehensive view of agriculture.
Evidence
“the Maha AgEx which is an open federated and consent driven architecture for data exchange it is helping us to bring diverse data sets together to get us a big picture.” [25].
Major discussion point
Central‑State Collaboration, Data Interoperability and Digital Public Infrastructure
Topics
Data governance | Artificial intelligence
Collaboration with MSSRF to place women’s rights at the centre of AI‑driven agriculture
Explanation
The speaker notes partnership with the M. S. Swaminathan Research Foundation to ensure that women’s rights and gender equity are integral to AI solutions in agriculture.
Evidence
“Swaminathan Research Foundation are working together on some of these issues in terms of how to bring women’s right in farming at the center stage.” [93]. “So we are happy to have support and assistance from MSSRF in that direction.” [95].
Major discussion point
Inclusion, Gender Equity and Empowerment of Women Farmers
Topics
Closing all digital divides | Human rights and the ethical dimensions of the information society
Open standards and interoperable architecture enable population‑scale transformation
Explanation
Rastogi highlights that open standards and interoperable systems have already demonstrated the ability to drive large‑scale change in agriculture, and that such approaches should be expanded.
Evidence
“These efforts have demonstrated how open standards and interoperable architecture can enable population -scale transformation that we are already seeing today.” [50].
Major discussion point
Central‑State Collaboration, Data Interoperability and Digital Public Infrastructure
Topics
Data governance | Artificial intelligence
AI for Agri‑2026 conference as a catalyst for global South knowledge exchange
Explanation
The panelist points to the upcoming AI for Agri‑2026 conference as a platform to deepen global collaboration, showcase use cases and foster South‑South knowledge sharing.
Evidence
“The session is also an important precursor to AI for Agri 2026 Global Conference, where we will continue these deliberations in greater operational depth with governments, investors, investors.” [107]. “AI for Agri conference is being held in Mumbai on 22nd and 23rd of February at Jio World Convention Centre.” [110].
Major discussion point
Role of Private Sector, Innovation, Financing and Global Partnerships
Topics
Artificial intelligence | International cooperation
Johannes Zutt
Speech speed
143 words per minute
Speech length
907 words
Speech time
377 seconds
“Thousand flowers bloom” approach encourages diverse farmer‑focused applications
Explanation
Zutt advocates for a permissive environment where many private innovators can experiment, allowing the most effective AI solutions for farmers to emerge organically.
Evidence
“And, you know, we can kind of let a thousand flowers bloom there and see what actually takes root.” [98]. “And, you know, that creativity will result in a number of different applications that will be aimed, in most cases, to help farmers overcome certain hurdles that they face.” [99].
Major discussion point
Role of Private Sector, Innovation, Financing and Global Partnerships
Topics
Artificial intelligence | The digital economy
World Bank can provide financing, technical assistance and truth‑testing for AI applications
Explanation
Zutt outlines the World Bank’s potential contributions: financing development, offering technical support, and conducting truth‑testing to ensure AI tools are credible and beneficial to farmers.
Evidence
“And then, you know, there are roles for institutions like my own, the World Bank Group, which can help to provide some of the financing that helps develop these applications, and also the foundational backbone for artificial intelligence.” [103]. “And we can also play a role at the advisory end, where we are helping to truth -test, if you like, the information that’s coming through different applications… to make sure that it’s actually providing information that is useful to the end beneficiary…” [104].
Major discussion point
Role of Private Sector, Innovation, Financing and Global Partnerships
Topics
Financial mechanisms | Artificial intelligence
Government responsibility for AI governance, interoperability and digital skilling
Explanation
Zutt notes that the government must lead on AI governance, ensure interoperable and accessible systems, and embed digital-skilling programs so that users can effectively engage with AI services.
Evidence
“The government’s responsibility is principally on foundations, things like the governance of AI, the interoperability, accessibility, obviously ensuring that educational programs include appropriate types of skilling in the use of digital services.” [6].
Major discussion point
Standards, Governance and Sustainable AI Deployments
Topics
Artificial intelligence | Capacity development
Dr. Soumya Swaminathan
Speech speed
173 words per minute
Speech length
1125 words
Speech time
388 seconds
Incorporate women’s data early to avoid algorithmic exclusion
Explanation
Swaminathan stresses that women’s data must be deliberately included from the outset because AI models are only as unbiased as the data they are trained on.
Evidence
“So I think it would be really important at the early stages itself to think about how women’s data can be incorporated because the algorithms are fed by the data we have.” [79].
Major discussion point
Inclusion, Gender Equity and Empowerment of Women Farmers
Topics
Closing all digital divides | Human rights and the ethical dimensions of the information society
AI should reduce drudgery for women and be co‑designed with them
Explanation
She argues that AI tools must be designed to lessen women’s workload, involve women in the design process, and ensure women’s representation on evaluation committees.
Evidence
“So I think there are a few points here, you know, to make sure particularly that women farmers are not left behind.” [81]. “…the women farmers and farmers in general must have a role they must be part of these committees that evaluate or make recommendations or make suggestions on improvement…” [82].
Major discussion point
Inclusion, Gender Equity and Empowerment of Women Farmers
Topics
Closing all digital divides | Human rights and the ethical dimensions of the information society
Design AI solutions “with women” not “for men”
Explanation
Swaminathan echoes the mantra that AI solutions should be built with women’s participation, not merely for them, to ensure relevance and equity.
Evidence
“with women farmers, not merely for them.” [83].
Major discussion point
Inclusion, Gender Equity and Empowerment of Women Farmers
Topics
Closing all digital divides | Human rights and the ethical dimensions of the information society
Human‑in‑the‑loop is essential for safe AI deployment
Explanation
She highlights that keeping humans in the decision loop is crucial to detect bias, prevent unintended harms and preserve employment opportunities in AI‑driven agriculture.
Evidence
“So I think Humans in the Loop is going to be important.” [119]. “And the last thing I would say is we still need humans in the loop.” [121].
Major discussion point
Standards, Governance and Sustainable AI Deployments
Topics
Artificial intelligence | Human rights and the ethical dimensions of the information society
Land‑title gaps leave most women farmers without ownership
Explanation
Swaminathan points out that the majority of women in India lack land titles, which can lead to their exclusion from AI‑driven services that rely on ownership data.
Evidence
“The first important fact is that women in India, the minority of them who have their name on the land document, so mostly it is in the man’s name… but that still means that, you know, three -fourths of them don’t have.” [84].
Major discussion point
Inclusion, Gender Equity and Empowerment of Women Farmers
Topics
Closing all digital divides | Human rights and the ethical dimensions of the information society
Shankar Maruwada
Speech speed
133 words per minute
Speech length
1259 words
Speech time
567 seconds
Open protocols (e.g., Beacon) enable interoperable networks and shared “rails” for AI models
Explanation
Maruwada describes how open protocols such as Beacon create interoperable networks that allow AI models and data services to be shared across states, fostering rapid diffusion while maintaining inclusion.
Evidence
“By creating interoperable networks based on open protocols like Beacon, by collaborating with each other, one of us is bringing in data, somebody is bringing in technology, somebody is bringing in policy, somebody is bringing in research.” [73]. “when I -CAR plugs into this network with its weather and pricing data that network makes it available to any state that wishes to turn on the supply… Shared Rails allows us to spread innovation diffuse it very quickly through society…” [74].
Major discussion point
Central‑State Collaboration, Data Interoperability and Digital Public Infrastructure
Topics
Data governance | Artificial intelligence
Standardizing AI ecosystems through open, networked architecture promotes sustainability
Explanation
He argues that adopting open, network‑style architectures and shared “rails” standardizes AI ecosystems, making them more sustainable, interoperable and easier to scale.
Evidence
“By creating interoperable networks based on open protocols like Beacon…” [73]. “…Shared Rails allows us to spread innovation diffuse it very quickly through society keeping in mind both inclusion and inclusion and inclusion and rewarding innovation because innovation has to be rewarded.” [74].
Major discussion point
Standards, Governance and Sustainable AI Deployments
Topics
Artificial intelligence | Data governance
Collaborative vision for open, interoperable digital public infrastructure
Explanation
Maruwada emphasizes a collaborative approach where multiple stakeholders contribute data, technology, policy and research to build open, interoperable digital infrastructure for agriculture.
Evidence
“It’s a very collaborative vision.” [102].
Major discussion point
Standards, Governance and Sustainable AI Deployments
Topics
Information and communication technologies for development | Artificial intelligence
Agreements
Agreement points
AI systems must be designed with inclusion and accessibility from the beginning, not as an afterthought
Speakers
– Dr. Soumya Swaminathan
– Shankar Maruwada
Arguments
AI systems must be designed with women farmers from the beginning, not as an afterthought, especially given agriculture’s increasing feminization
Systems must be designed for the most marginalized users, including illiterate farmers with basic phones, to ensure true inclusion
Summary
Both speakers emphasize that inclusive design must be built into AI systems from the start, with Swaminathan focusing on women farmers and Maruwada on the most marginalized users including illiterate farmers with basic technology
Topics
Closing all digital divides | Artificial intelligence | Human rights and the ethical dimensions of the information society
Open, interoperable systems and digital public infrastructure are essential for scaling AI solutions
Speakers
– Devendra Fadnavis
– Dr. Devesh Chaturvedi
– Shankar Maruwada
– Vikas Chandra Rastogi
Arguments
AI must be built on trusted data, ethical governance, and public accountability to achieve scale and farmer adoption
Digital public infrastructure with farmer IDs and integrated AI systems like Bharatvistar can eliminate digital red tapism and provide tailored advisories
Open interoperable systems based on DPI principles allow collaborative innovation and rapid diffusion of AI solutions across states
MahaAgEx creates open federated architecture for data exchange to integrate diverse datasets for comprehensive agricultural insights
Summary
All speakers agree that open, interoperable digital infrastructure is crucial for scaling AI solutions effectively, eliminating fragmentation, and enabling collaborative innovation across different stakeholders
Topics
Artificial intelligence | Information and communication technologies for development | The enabling environment for digital development | Data governance
Government and private sector have complementary roles in AI development and deployment
Speakers
– Johannes Zutt
– Devendra Fadnavis
Arguments
AI applications require creativity from private sector to develop specific solutions for farmer constraints, with government providing foundational support
AI can provide hyperlocalized advisories, predictive analytics, and transparent supply chains to transform agriculture at scale
Summary
Both speakers recognize that successful AI deployment requires government to provide foundational infrastructure and governance while private sector brings innovation and creativity to develop specific applications
Topics
Artificial intelligence | The enabling environment for digital development | Financial mechanisms
Trust, governance, and accountability are fundamental for AI adoption and success
Speakers
– Devendra Fadnavis
– Johannes Zutt
– Dr. Soumya Swaminathan
Arguments
AI must be built on trusted data, ethical governance, and public accountability to achieve scale and farmer adoption
Government responsibility includes AI governance, interoperability, accessibility, and ensuring research-backed credible advisories
Ongoing evaluation with feedback loops and farmer voices in decision-making committees is essential for responsible AI deployment
Summary
All three speakers emphasize that trust, proper governance, and accountability mechanisms are essential for AI systems to be adopted by farmers and achieve meaningful impact
Topics
Artificial intelligence | Human rights and the ethical dimensions of the information society | The enabling environment for digital development
Similar viewpoints
Both government leaders highlight the success of AI platforms like Mahavistar and Bharatvistar in serving millions of farmers, demonstrating that farmers are ready to adopt AI when it’s properly designed and integrated with digital public infrastructure
Speakers
– Devendra Fadnavis
– Dr. Devesh Chaturvedi
Arguments
Mahavistar platform serves 2.5 million farmers with multilingual advisories and demonstrates farmer readiness for AI when designed appropriately
Digital public infrastructure with farmer IDs and integrated AI systems like Bharatvistar can eliminate digital red tapism and provide tailored advisories
Topics
Artificial intelligence | Information and communication technologies for development | Social and economic development
Both speakers recognize India’s unique position to lead AI development globally due to its scale, diversity, and experience with digital public infrastructure, enabling it to create solutions that can benefit other developing countries
Speakers
– Johannes Zutt
– Shankar Maruwada
Arguments
India’s scale, diversity, and DPI experience positions it to lead AI development for developing countries and South-South learning
Open interoperable systems based on DPI principles allow collaborative innovation and rapid diffusion of AI solutions across states
Topics
Artificial intelligence | Information and communication technologies for development | Capacity development
Both speakers emphasize the need to address women farmers’ specific challenges, including lack of documentation and the need to reduce their workload, ensuring AI systems are designed to include and benefit women farmers specifically
Speakers
– Dr. Soumya Swaminathan
– Vikas Chandra Rastogi
Arguments
Women farmers often lack land ownership documentation, so AI systems risk excluding them if based only on publicly available data
AI solutions should reduce drudgery and workload on women farmers, particularly in remote and tribal areas where traditional methods persist
Topics
Closing all digital divides | Human rights and the ethical dimensions of the information society | Social and economic development
Unexpected consensus
Human-in-the-loop approach rather than complete automation
Speakers
– Dr. Soumya Swaminathan
– Shankar Maruwada
Arguments
Human-in-the-loop approaches remain important rather than complete automation, maintaining employment while augmenting traditional knowledge
Systems must be designed for the most marginalized users, including illiterate farmers with basic phones, to ensure true inclusion
Explanation
Despite the focus on advanced AI technology, there’s unexpected consensus that human involvement remains crucial. This is significant because it shows recognition that technology should augment rather than replace human knowledge and employment, especially in a country like India where employment is critical
Topics
Artificial intelligence | Social and economic development | Capacity development
Need for continuous evaluation and iteration rather than perfect deployment
Speakers
– Dr. Soumya Swaminathan
– Shankar Maruwada
Arguments
Ongoing evaluation with feedback loops and farmer voices in decision-making committees is essential for responsible AI deployment
Open interoperable systems based on DPI principles allow collaborative innovation and rapid diffusion of AI solutions across states
Explanation
There’s unexpected consensus that AI systems should be deployed as minimum viable products and improved iteratively, rather than waiting for perfect solutions. This represents a shift from traditional technology deployment approaches and shows pragmatic understanding of AI development
Topics
Artificial intelligence | Monitoring and measurement | Human rights and the ethical dimensions of the information society
Overall assessment
Summary
There is strong consensus among all speakers on the fundamental principles of AI deployment in agriculture: the need for inclusive design, open interoperable systems, proper governance and trust mechanisms, and collaborative approaches between government and private sector. All speakers agree that India’s digital public infrastructure experience positions it well to lead responsible AI deployment at scale.
Consensus level
High level of consensus with complementary perspectives rather than disagreements. The speakers represent different sectors (government, international development, research, technology) but share aligned visions for responsible AI deployment. This strong consensus suggests favorable conditions for implementing the discussed AI initiatives and policies, with clear agreement on both technical approaches and governance principles.
Differences
Different viewpoints
Role of human involvement vs automation in AI systems
Speakers
– Dr. Soumya Swaminathan
– Shankar Maruwada
Arguments
Human-in-the-loop approaches remain important rather than complete automation, maintaining employment while augmenting traditional knowledge
Systems must be designed for the most marginalized users, including illiterate farmers with basic phones, to ensure true inclusion
Summary
Dr. Swaminathan emphasizes the continued importance of human involvement and warns against complete automation, while Maruwada focuses on designing systems that can work independently for the most marginalized users with minimal human intervention
Topics
Artificial intelligence | Capacity development | Human rights and the ethical dimensions of the information society
Approach to AI system design and deployment
Speakers
– Dr. Soumya Swaminathan
– Dr. Devesh Chaturvedi
Arguments
Ongoing evaluation with feedback loops and farmer voices in decision-making committees is essential for responsible AI deployment
Digital public infrastructure with farmer IDs and integrated AI systems like Bharatvistar can eliminate digital red tapism and provide tailored advisories
Summary
Dr. Swaminathan advocates for careful evaluation and farmer involvement in decision-making before wider deployment, while Dr. Chaturvedi emphasizes rapid deployment of integrated systems to eliminate bureaucratic barriers
Topics
Artificial intelligence | Human rights and the ethical dimensions of the information society | The enabling environment for digital development
Unexpected differences
Pace and approach to AI deployment in agriculture
Speakers
– Dr. Soumya Swaminathan
– Dr. Devesh Chaturvedi
– Shankar Maruwada
Arguments
Ongoing evaluation with feedback loops and farmer voices in decision-making committees is essential for responsible AI deployment
Digital public infrastructure with farmer IDs and integrated AI systems like Bharatvistar can eliminate digital red tapism and provide tailored advisories
Open interoperable systems based on DPI principles allow collaborative innovation and rapid diffusion of AI solutions across states
Explanation
Unexpected because all speakers support AI in agriculture, but they differ significantly on deployment approach – Swaminathan advocates for careful evaluation and farmer involvement, while Chaturvedi and Maruwada support rapid deployment with iterative improvement. This disagreement on methodology could impact the success of AI implementation
Topics
Artificial intelligence | Human rights and the ethical dimensions of the information society | The enabling environment for digital development
Overall assessment
Summary
The discussion shows broad consensus on AI’s potential for agriculture but reveals subtle yet significant disagreements on implementation approaches, particularly regarding the balance between rapid deployment and careful evaluation, the role of human involvement versus automation, and specific inclusion strategies
Disagreement level
Low to moderate disagreement level with high implications – while speakers agree on goals, their different approaches to implementation could lead to significantly different outcomes in terms of inclusion, effectiveness, and farmer adoption of AI systems
Partial agreements
Partial agreements
Both agree on the need for responsible AI deployment and inclusion, but Fadnavis focuses on general trust and governance principles while Swaminathan specifically emphasizes gender inclusion and women farmers’ needs from the design stage
Speakers
– Devendra Fadnavis
– Dr. Soumya Swaminathan
Arguments
AI must be built on trusted data, ethical governance, and public accountability to achieve scale and farmer adoption
AI systems must be designed with women farmers from the beginning, not as an afterthought, especially given agriculture’s increasing feminization
Topics
Artificial intelligence | Human rights and the ethical dimensions of the information society | Closing all digital divides
Both support collaborative approaches between public and private sectors, but Zutt emphasizes distinct roles for government (foundations) and private sector (creativity), while Maruwada focuses on open collaborative networks where all stakeholders contribute different elements
Speakers
– Johannes Zutt
– Shankar Maruwada
Arguments
AI applications require creativity from private sector to develop specific solutions for farmer constraints, with government providing foundational support
Open interoperable systems based on DPI principles allow collaborative innovation and rapid diffusion of AI solutions across states
Topics
Artificial intelligence | The enabling environment for digital development | Financial mechanisms
Similar viewpoints
Both government leaders highlight the success of AI platforms like Mahavistar and Bharatvistar in serving millions of farmers, demonstrating that farmers are ready to adopt AI when it’s properly designed and integrated with digital public infrastructure
Speakers
– Devendra Fadnavis
– Dr. Devesh Chaturvedi
Arguments
Mahavistar platform serves 2.5 million farmers with multilingual advisories and demonstrates farmer readiness for AI when designed appropriately
Digital public infrastructure with farmer IDs and integrated AI systems like Bharatvistar can eliminate digital red tapism and provide tailored advisories
Topics
Artificial intelligence | Information and communication technologies for development | Social and economic development
Both speakers recognize India’s unique position to lead AI development globally due to its scale, diversity, and experience with digital public infrastructure, enabling it to create solutions that can benefit other developing countries
Speakers
– Johannes Zutt
– Shankar Maruwada
Arguments
India’s scale, diversity, and DPI experience positions it to lead AI development for developing countries and South-South learning
Open interoperable systems based on DPI principles allow collaborative innovation and rapid diffusion of AI solutions across states
Topics
Artificial intelligence | Information and communication technologies for development | Capacity development
Both speakers emphasize the need to address women farmers’ specific challenges, including lack of documentation and the need to reduce their workload, ensuring AI systems are designed to include and benefit women farmers specifically
Speakers
– Dr. Soumya Swaminathan
– Vikas Chandra Rastogi
Arguments
Women farmers often lack land ownership documentation, so AI systems risk excluding them if based only on publicly available data
AI solutions should reduce drudgery and workload on women farmers, particularly in remote and tribal areas where traditional methods persist
Topics
Closing all digital divides | Human rights and the ethical dimensions of the information society | Social and economic development
Takeaways
Key takeaways
AI integration in agriculture must be built on open, interoperable digital public infrastructure (DPI) principles to achieve population-scale impact while ensuring inclusion and trust
Maharashtra’s Maha Agri AI Policy 2025-2029 demonstrates successful AI implementation with Mahavistar serving 2.5 million farmers, proving farmers are ready for AI when designed appropriately
Women farmers must be centered in AI system design from the beginning, not as an afterthought, especially given agriculture’s increasing feminization and their lack of land ownership documentation
AI systems should move from fragmented digital red tapism to integrated platforms providing hyperlocalized, multilingual advisories based on farmer IDs and comprehensive data integration
India’s scale, diversity, and DPI experience positions it uniquely to lead responsible AI development for developing countries and facilitate South-South knowledge exchange
Government, private sector, and development institutions must collaborate with distinct roles – government providing foundational infrastructure and governance, private sector driving innovation, and development partners offering financing and advisory support
AI deployment must include ongoing evaluation, feedback loops, and farmer voices in decision-making to ensure responsible scaling and address unintended biases or exclusions
Human-in-the-loop approaches remain essential rather than complete automation, maintaining employment while augmenting traditional knowledge and reducing farmer drudgery
Resolutions and action items
Launch AI for Agri 2026 Global Conference in Mumbai on February 22-23 at Jio World Convention Centre to continue operational discussions with governments, investors, and innovators
Expand Bharatvistar to all Bhashini-related languages within 3-6 months from current English and Hindi availability
Achieve saturation of farmer IDs and digital crop survey across states, with close to 9 crore farmer IDs already developed
Implement very targeted, tailored advisories based on farmer ID integration with digital public infrastructure within 3-6 months
Expand predictive models for market and weather situations to improve farmer decision-making and increase productivity while reducing costs
Develop traceability digital public infrastructure as a replicable public infrastructure model for India and the global south
Create global partnership opportunities for venture capital funds, impact investors, multilateral development banks, and corporate innovation arms to partner with Maharashtra
Government of Maharashtra and Dr. M.S. Swaminathan Research Foundation to collaborate on women’s rights in farming, bio-happiness creation, and nutritional security initiatives
XTEP Foundation’s declaration to create 100 AI diffusion pathways by 2030 across different sectors, countries and continents
Unresolved issues
How to effectively incorporate women farmers’ data into AI systems when majority lack land ownership documentation
Specific mechanisms for ensuring AI systems reduce drudgery for women farmers in remote and tribal areas using traditional farming methods
Detailed governance frameworks and institutional safeguards needed for responsible AI scaling while maintaining scientific integrity
Cost and connectivity challenges for farmers with limited assets and basic smartphones in remote areas
Specific evaluation methodologies and success indicators for measuring AI impact on inclusion and farmer outcomes
How to balance complete automation versus human-in-the-loop approaches while maintaining employment in rural areas
Detailed technical specifications for interoperability standards across different state AI systems and private sector applications
Funding mechanisms and sustainability models for long-term AI infrastructure maintenance and continuous improvement
Suggested compromises
AI should augment rather than replace traditional extension services and human knowledge systems, working as an additional resource rather than complete replacement
Balance between government-led foundational infrastructure and private sector innovation through collaborative open networks that reward innovation while ensuring inclusion
Iterative deployment approach where AI systems start with minimum viable products and evolve over time rather than waiting for perfect technology before deployment
Shared rails approach allowing states flexibility to choose and customize AI applications while maintaining interoperability through common protocols
Collaborative investment model where different stakeholders contribute their strengths – data, technology, policy, research – rather than siloed development
Thought provoking comments
AI is not a magic. As Honorable PM said in his inaugural session, AI must be built on trusted data, ethical governance, and public accountability. Without trust, scale will not happen.
Speaker
Devendra Fadnavis
Reason
This comment is profoundly insightful because it cuts through the AI hype and addresses the fundamental challenge of responsible AI deployment at scale. It acknowledges that technology alone is insufficient without proper governance frameworks and public trust.
Impact
This comment set the tone for the entire discussion by establishing that the conversation would focus on practical implementation challenges rather than just technological possibilities. It influenced subsequent speakers to address governance, trust, and institutional frameworks throughout their responses.
So basically a farmer who has to avail so many services was, we felt that he or she was getting lost in which app to use… Sometimes it becomes more difficult to avail the services through online systems… So the whole idea was that once we have this AI-based system, we have a same platform for different applications and different advisories at a click of the button.
Speaker
Dr. Devesh Chaturvedi
Reason
This observation is thought-provoking because it identifies a critical paradox in digital transformation – that digitization can create ‘digital red tapism’ that’s worse than the original bureaucratic problems it was meant to solve. It reframes AI not just as advanced technology, but as a solution to user experience fragmentation.
Impact
This comment shifted the discussion from technical capabilities to user-centric design principles. It influenced the conversation to focus on interoperability and unified platforms, which became a recurring theme throughout the panel discussion.
So I would say that, you know, one of the benchmarks that I would look at is, is it reducing the drudgery and the workload on women farmers? Is AI helping to do that? So I think we also need to look at certain indicators for success.
Speaker
Dr. Soumya Swaminathan
Reason
This comment is insightful because it introduces concrete, measurable criteria for AI success that goes beyond productivity metrics to include social impact. It challenges the discussion to think about AI’s role in addressing gender equity and labor conditions.
Impact
This comment introduced a critical evaluation framework that influenced the discussion to consider not just what AI can do, but what it should do. It elevated the conversation from technical implementation to social responsibility and outcome measurement.
A system that operates basically on publicly available data will then leave out those whose data sets are not available… And so all of these advisories may be very suitable for a man who’s operating a tractor on a farm, but not at all relevant for a woman who’s still working with outdated instruments.
Speaker
Dr. Soumya Swaminathan
Reason
This observation is particularly thought-provoking because it exposes a fundamental bias in AI systems – that they can perpetuate and amplify existing inequalities through data gaps. It challenges the assumption that AI is inherently neutral or inclusive.
Impact
This comment introduced a critical perspective on algorithmic bias that influenced the discussion to consider inclusion as a design principle rather than an afterthought. It prompted deeper consideration of how data collection and system design can either include or exclude marginalized groups.
Unlike the technologies of the past where you perfect the technology and then deploy AI, you deploy something minimum to start and then evolution, models get better, data gets better, usage gets better. And then it gets better and better over time.
Speaker
Shankar Maruwada
Reason
This insight is thought-provoking because it articulates a fundamental shift in technology deployment philosophy – from perfectionist approaches to iterative, learning-based systems. It recognizes AI’s unique characteristic of improving through use.
Impact
This comment reframed the entire approach to AI implementation, shifting the discussion from concerns about readiness and perfection to embracing experimentation and continuous improvement. It influenced the conversation to focus on building adaptive systems rather than static solutions.
When Edmund Hillary climbed Mount Everest, he made a lot of people believe it is possible. When Mahavistar was launched, it made the country believe that it is possible to make AI serve the farmer.
Speaker
Shankar Maruwada
Reason
This analogy is insightful because it positions technological achievements as proof-of-concept that inspire broader adoption and innovation. It recognizes the psychological and social dimensions of technology diffusion beyond just technical capabilities.
Impact
This comment provided a powerful concluding framework that positioned the current initiatives as pathbreaking efforts with broader inspirational value. It elevated the discussion from local implementation to global leadership and responsibility.
Overall assessment
These key comments fundamentally shaped the discussion by moving it beyond technical specifications to address the deeper challenges of responsible AI deployment at scale. The conversation evolved from initial enthusiasm about AI capabilities to a more nuanced understanding of implementation challenges, including user experience design, social inclusion, algorithmic bias, and governance frameworks. The most impactful comments introduced critical evaluation criteria and philosophical frameworks that elevated the discussion from ‘how to build AI systems’ to ‘how to build AI systems that serve society equitably and effectively.’ The speakers built upon each other’s insights, creating a comprehensive framework for AI in agriculture that balances innovation with responsibility, scale with inclusion, and technological capability with human-centered design.
Follow-up questions
How to ensure that AI systems incorporate women farmers’ data when majority of land documents are not in women’s names?
Speaker
Dr. Soumya Swaminathan
Explanation
This is critical for inclusive AI development as algorithms are fed by available data, and excluding women farmers’ data could make AI advisories irrelevant for their specific needs and farming practices
How to develop evaluation frameworks and clinical trial-like methodologies for AI agricultural applications?
Speaker
Dr. Soumya Swaminathan
Explanation
As a medical researcher, she emphasized the need for rigorous evaluation of AI products similar to clinical trials to examine data, evidence, inherent biases, and unanticipated risks before wider rollout
How to ensure farmers and women farmers are included in committees that evaluate and make recommendations for AI systems?
Speaker
Dr. Soumya Swaminathan
Explanation
She noted the absence of actual farmers in the discussion and emphasized that those for whom AI systems are developed must have a voice in the evaluation and improvement process
How to structure programs and technical assistance models that provide just-in-time support to governments for AI experimentation and scaling?
Speaker
Vikas Chandra Rastogi
Explanation
This addresses the need for development partnerships to adapt to rapidly evolving AI technologies and provide agile, responsive support to enable responsible AI deployment
How to ensure AI reduces drudgery and workload specifically for women farmers in remote and tribal areas?
Speaker
Dr. Soumya Swaminathan
Explanation
This is important as women in remote areas often use traditional methods and tools, and AI should address their specific constraints rather than being designed primarily for mechanized farming
How to maintain the balance between AI automation and human employment in agriculture?
Speaker
Dr. Soumya Swaminathan
Explanation
She emphasized the need for ‘humans in the loop’ approach, noting that in a country like India, employment generation is crucial and complete automation may not be the solution
How to create 100 AI diffusion pathways across different sectors and countries by 2030?
Speaker
Shankar Maruwada
Explanation
This represents XTEP Foundation’s ambitious vision to create multiple collaborative pathways for safe AI impact at scale, requiring significant coordination and partnership across sectors and geographies
How to ensure AI advisories are relevant for diverse farming practices, especially traditional methods used by women in millet cultivation?
Speaker
Dr. Soumya Swaminathan
Explanation
This addresses the need for AI systems to be inclusive of traditional and subsistence farming practices, particularly those predominantly managed by women farmers
Disclaimer: This is not an official session record. DiploAI generates these resources from audiovisual recordings, and they are presented as-is, including potential errors. Due to logistical challenges, such as discrepancies in audio/video or transcripts, names may be misspelled. We strive for accuracy to the best of our ability.
Related event
