AI Use Cases in Farmer Co-operatives

This document describes several examples of how Artificial Intelligence and connected data are being used to support agriculture in India, accompanied by Patrick Shulist’s analysis of the potential benefits and challenges that these and other applications of AI technology may present for farmers in lower income countries.

Low-cost Sowing Advisory

The date of sowing a crop is critical in determining the quality and yield of a harvest. In India, the sowing date is determined by age old ancestral traditions and practices. However, with more erratic weather changes within the country, the monsoon season is not as uniform and easily predictable as before. Unusually wet or dry spells that are not accounted for during sowing can negatively affect harvests and cause huge losses to farmers. 

As such, a pilot program from ICRISAT (International Crops Research Institute for the Semi-Arid Tropics) and Microsoft seeks to mitigate such losses through the help of an AI driven sowing advisory. The AI driven system is fed with historic climate data spanning over 30 years of the region in question. To determine the optimal sowing period, a Moisture Adequacy Index (MAI) is calculated in real-time. MAI is the standardized measure used for assessing the degree of adequacy of rainfall and soil moisture to meet the potential water requirement of crops. Along with historic data and future weather forecasting models, the AI technology guides farmers on which day/week is ideal to sow crops. 

The sowing advice is disseminated to farmers via a simple SMS on their mobile phones. The SMS also contains additional advice such as the recommended sowing depth, the amount of water to be used, as well as the recommended amount of fertilizers. The program has shown great results in the Kurnool district of Andhra Pradesh (where it was piloted), with farmers reporting a 30% increase in harvest yield. 

The success of future applicability of this intervention lies in its low cost. For low-literate and low-income farmers, simple SMS data insights are easily understood and accessible. Farmers do not have to invest in expensive phones or have to learn the functionality of complicated apps. Furthermore, this intervention does not require capital expenditure on hardware such as farm sensors. Within co-operatives, the intervention can be tweaked to include a more insightful ‘dashboard’ with more information at a co-operative level – i.e with the management of the co-operative rather than with each individual – in order to make more complex insights affordable.

AI Chat Bots

There is a big need for access to real time knowledge and insights within the Indian farming community. The ability to have questions answered, problems addressed, and plant diseases diagnosed is integral in making farming more efficient and effective. Currently, the channels to have these addressed are very long and in some cases costly. Farmers have to wait for long periods to meet with an expert, and often just resort to age old ancestral knowledge^{1, 2}.

To address this, Agri Tech companies are developing AI driven chat bots to provide farmers with instant advice on various topics and issues. For instance, Digital Green, in collaboration with ColouredCow, has developed a chat bot to provide farmers with customised notifications and videos on a real-time basis to help manage their crops more efficiently. This chatbot is specifically developed for chilli producers in Andhra Pradesh to receive real time information on the curled leaf disease that commonly effects the chilli crop. Farmers may ask questions regarding the prevention or treatment of the disease and receive customised answers based on the stage of the crop growth, the advancement of the disease and the region they are based in.

This intervention also helps gather and aggregate data that can be used by farmer cooperatives. The software provides interactive graphs that include data on the crops grown by the farmers, the land holdings, as well as personal data of the farmers themselves.

While this is an extremely important and helpful AI based intervention, it would require a tremendous amount of data to train the algorithm and provide customised and relevant answers. As such, this intervention may be suited for specific cases that would help reduce the amount of training and data required – such as that of a particular disease on chilli farms.

Agri Drones – Krishi Viman

In 2020, The Indian Government approved the use of UAV Drones (un-manned aerial vehicle) for agricultural activities. Since then, the Ministry of Agriculture has been heavily pushing the application of Drones for precision farming by budgeting for the production of Indian-made UAV’s as well as developing a robust regulation around it. 

In lieu of the increasing support for UAV’s in agriculture, WOW Go Green, an Indian based Agro Tech company has developed the Krishi Viman – a smart Drone – with the support of the government. The drone’s main functionality is spraying pesticides on crops based on thermal imaging, surveillance and mapping. As such, The Krishi Viman aims to reduce manual labour, lower health hazards, and improve efficiency.

The benefits of the Krishi Viman are not limited to merely its functionality. To support the adoption and use of this technology, the organisation also provides certain support services. Considering the expenses involved in procuring this technology, the farmers are given loan and credit support to help allow easier access to the Krishi Viman. They are also provided with a pilot training program, insurance support and a 24/7 helpline for queries. 

While such support services help in making the Krishi Viman more accessible to low-literate and low-income farmers, it is still a very expensive technology. An immense amount of resources may be needed to develop the necessary infrastructure and scale the Krishi Viman to an extent where it is easily available to small-scale farmers. In the meantime, this technology may be well-suited for farmer cooperatives, where the cost of such equipment is distributed among farming communities, and the ownership of the technology comes under the co-operative as a whole.

AI-based Price Predictions

Price prediction is an important data insight for Indian farmers, especially in an agriculture market that is ridden with high price volatility. Very often, farmers that are not aware of upcoming price and market trends do not plan their harvest and sales effectively. They are forced to sell their produce at prices that don’t fetch them any profits. As such, in order to mitigate such risks and protect farmers, there is a dire need for accurate and hyper-local price predictions. The predictive capacity of AI can be extremely helpful here. 

IBM, in collaboration with the Karnataka Agricultural Price Commission (KAPC) is developing an advanced price forecasting system—a dashboard leveraging IBM’s Watson Decision Platform for Agriculture. The initiative is being launched in the districts of Kolar, Chikkaballapur and Belgavi to predict market prices of tomato and maize. The AI based forecasting system uses data from satellite imagery and weather forecasts to assess crop acreage, crop health and crop yield on a real-time basis. It then blends this with historic and current Pricing data from neighbouring markets to forecast crop prices.

In another initiative, a team of researchers from Pennsylvania State University developed an algorithm to predict crop prices in India. This algorithm was based on analysing India’s crop market prices and volumes for over 11 years. The algorithm advises farmers on where and when to sell their produce such that they can maximize their profits. For instance, instead of selling produce the very next day after harvesting, this technology may advise the farmers to wait a few days, or travel to a different market to obtain a better selling price. 

Such an intervention would be well suited in co-operative settings, where the co-operative can help the service providers by collecting and sharing aggregated price and volume data from its farmers. This would help feed data to the system to obtain hyper-local price predictions. Moreover, it would be accessible to low-income and low-literate farmers as it does not require them to handle any expensive or complicated machinery or software. The suggestions can be disseminated via SMS through the service providers themselves, or by the farmer co-operative management. 

Other Use-Cases

There are plenty of other use cases that can be used by low-income and low-literate farmer co-operatives, especially with an investment in capacity building and infrastructure development to make them more accessible to the farmers.

Soil Health Monitoring

Farm based tools can feed important data to AI algorithms to help ascertain vital farm related information and insights, such as soil health. Remote ground sensors or hand-held tools can collect various soil-related indicators such as soil moisture, micro and macro nutrients and an overall profile of the soil present on a farmer’s land. Such data can be fed into an AI-based algorithm to get insights on how healthy the soil is, what fertilizer must be used, how often it must be watered, etc., such that it facilitates a healthy crop yield. 

Pest/Disease prediction, identification and mitigation

Imaging tools such as drones, on-field cameras, thermal cameras, etc., can help identify and diagnose crop diseases based on image analysis through AI-backed software. Timely diagnosis and communication of such diseases will be vital in minimizing crop losses. Additionally, farmers can upload pictures of plants or pests found on their farmland to Applications that provide real time identification, diagnosis and advice on managing possible diseases or pest infestations. Real-time video feeds can also help in identifying animal breaches in the farm. Accordingly, alerts to farmers or an alarm system can be triggered to prevent any crop destruction. 

Smart Irrigation Management

In India, farm land is mostly dependent on the monsoons for the majority of its water source.  The efficiency of irrigation within farms is low, and in the case of an erratic monsoon there is a high risk of crop loss owing to sub-par irrigation forecasting. AI-based software can provide forecasts for tracking water availability from surface, ground and soil moisture. Based on these forecasts, the software can ascertain water surplus and deficit regions and aid in village-level water budgeting to make irrigation more efficient and effective.

Smart Insurance

Precision agriculture, alongside machine learning, has the potential to reduce agricultural insurance premiums by defining risks and improving risk assessment tools, and allows farms to move more quickly to prevent crop losses. Moreover, the data collected and stored in farm tools and software can help in verifying insurance claims of the farmers as well as the companies.

Accessibility of Credits and Loans

A lot of farmers struggle to secure loans based on their inability to provide collateral against such loans. Additionally, it is difficult for lenders to ascertain an individual’s credit-worthiness owing to a lack of data on their net worth or previous financial activities. Data collected through AI tools and software such as soil health monitoring, weather forecasts and yield forecasts can now be used by lenders as a means to ascertain a farmer’s earnings capacity and hence their financial health and creditworthiness. Such forecast data can also be used to ascertain how much credit may be required for farming related investments and what value they can generate.

Footnotes

1. For an alternative perspective on the importance of traditional practices and their role in sustainable agriculture and communities as understood by the farmer-owners of two SEWA agriculture cooperatives, see Kumar, Ranjitha, Anuradha Ganapathy and Sakhi Shah (2022). “SEWA Cooperative Federation Baseline Report.”
2. SEWA’s Agricultural Knowledge Databank provides an alternative model that integrates traditional farming knowledge with the collection and analysis of data contributed by workers. See Basman, Antranig (2023). “Cooperatively developed agricultural practices.” Data Communities for Inclusion Governance Toolkit.