Lauding the role of AI in agriculture, particularly in improving the yield of crops and income of farmers, experts have expressed concern over te quality of advisories as the sector is not regulated and suggested that standards should be developed.
Addressing a session on Standards and Policy Pathways for Scaling AI-Enabled, Farmer -Centric Agrifood Systems in New Delhi, Santanu Chaudhary, former director of IIT-Jodhpur, said: “If AI is driving the public decision-making regarding agriculture and agricultural advice, who will be accountable regarding the quality of the advice? This is fundamentally important because it impacts livelihoods of many farmers.”
Coinciding with the AI event, Union Agriculture Minister Shivraj Singh Chouhan is scheduled to launch the newly announced Bharat-VISTAAR scheme on Tuesday.
Registration process
Chaudhary said there are large number of start-ups in the US, where government initiatives focus on enabling interventions and driving scale through AI. But in India, he said something that is extremely important to address today is to bring AI in the farm sector with appropriate accountability mechanisms.
Elaborating further, he said “it must be clearly monitored, should be traceable and should be challenged”. This contestable advice is critical because it ensures that the guidance provided is used solely for the intended purpose and not influenced by business interests or other unethical interventions, he said.
As AI now comes in models, there has to be a process to register the models, he noted. “A registration should have a process and protocol defined. In the absence of any protocol, if one says that its AI ensures 99 per cent accuracy, it doesn’t mean anything because that is with respect to the data collected and how general is the performance across multiple stakeholders, multiple scenarios, he said.
Problem of drift
Since farm land in India is fragmented, soil property varies depending on the land’s usage history. As a result, AI models that make predictions based on land data may experience drift when applied from one plot to another, Chaudhary said.
Highlighting that International Telecommunication Union has a very robust standardisation on digital agriculture and standards, ITU’s Atusko Okuda said it has over 200 AI-related standards that are already approved and 200 more are in the pipeline.
She also said these standards are critical for the democratisation of AI in digital agriculture because not all countries are equipped to start from zero. “So, when you have standards, it is easier for countries to kickstart from where the standard ends,” she said, adding ITU has also standards for very specific areas of agriculture such as smart livestock farming based on IoT.
Raghu Chaliganti of Fraunhofer HHI, said India has a large number of small farmers at about 130 million, and reaching them is a major challenge. “How to bring the innovations and this AI technology to their doorsteps? Farmers want input cost to reduce and production to increase and better price for the produce,” he said adding that AI is promoted as a solution that can minimise the input cost and maximise productivity.
Published on February 16, 2026