Table of Contents
This was a multi-part series focusing on the applications of AI that I wrote for the Insprit AI ambassador program. I’ve compiled them here into one post.
Part One: Why agriculture faces challenges today, and what those are #
Being one of the significant challenges facing humanity in the present day as we look towards the future, food supply and agriculture are ripe to be transformed by continuous innovations and progress in artificial intelligence. If we’re to produce 70% more food than we do today–which we’ll need to sustain the increasing demands of a growing population–AI will have a role in helping us get there.
In this first instalment of a three-part series, we’ll take a deep dive into the problems faced by the agriculture industry today — and a sneak peek at some of the groundbreaking solutions to these modern-day challenges.
Is this the right place to be putting resources? #
Agriculture isn’t a small industry by any means — nor is it inconsequential. Apart from employing 1.3 billion people worldwide, the UN’s Sustainable Development Goals goal of zero hunger by 2030 will be improbable without significantly improved agricultural output and efficiency across the globe. As much as progress has occurred in agricultural technologies in the recent past, there remains much-untapped potential for improvements and disruption.
In this post, we’ll take a deep dive into those problems.
1 - Ineffective resource allocation #
Agriculture is notorious for not being frugal in its usage of precious resources like water — approximately 70% of all water usage worldwide is due to it, and of that, 60% is wasted. As freshwater becomes increasingly valuable, agriculture sees a need to explore new avenues for efficient water use as much as, if not more than, any other sector. Water is not the only resource whose usage leaves room to be optimised. Chemicals and pesticides are, in many cases, indiscriminately used in a non-data-driven manner for optimising when and where to use them, thus leading to much waste. Every year in the US alone, a billion pounds of pesticide are used, and many water supplies globally are contaminated with pesticides and other agricultural products. Put together, agriculture emerges as a significant factor in water pollution.
2 - A shrinking and ageing labour force #
The agricultural labour force is shrinking and ageing, as finding someone or something to run such farms is becoming more complex. The average Indian farmer is 51 years old; the average American is 58. A rapidly increasing population demands increasing agricultural output, but the farmers who supply that food are ageing, which bids poorly for the workforce’s future. The percentage of the population employed as part of the agricultural workforce is also shrinking — almost 40% of the working population in 2000 was employed in agriculture; in 2020, that figure is only 26%, according to an ILO estimate. Youth worldwide seem to be progressively less interested in agriculture as other fields of work are becoming increasingly popular.
External factors #
3 - Climatic and weather-related events #
Climate and weather-related events can ravage farmers. According to an FAO report, the global annual average economic damage to agriculture between 2004 to 2014 was 100 billion USD. Farmers bear the brunt of such events: floods, storms, droughts, storms, hurricanes, cyclones, or wildfires. In particular, crops are most affected by floods and storms.
4 - Pest Control and Crop Disease #
Globally, pests and crop disease reduce the output of five major food crops by 10 to 40 per cent.  Over millennia, pathogens and pests have evolved along with human crops to be more effective. Invasive insects alone cost $70 billion a year worldwide. Put together with crop diseases, they pose a significant threat to the global food supply, environment, and human health in more significant part due to the agricultural sector’s inability to combat them.
5 - Financial Burden #
Several farmers worldwide cannot procure financing for near-necessary purchases such as farm inputs and harvesting crops, particularly in developing countries such as Kenya. Many farmers are also burdened by significant debt if they can procure loans. Furthermore, several farmers do not keep financial records, making obtaining the historical performance of their farms and financial projections difficult.
What’s Next? #
These are just some of the challenges the increasingly vital agricultural sector faces. In the following post, we’ll look at possible solutions to the issues described here.
Parts 2 and 3: AI-Driven Solutions for Agriculture Problems #
In the previous article of this series, we looked at several of the problems agriculture faces as it tries to scale to adapt to 21st-century demand. Now, over the following two articles, we’ll take a deep dive into the innovative solutions that have manifested to resolve these pivotal issues.
Problem: Resource Allocation #
As we saw in the last article, approximately 70% of all water usage worldwide is due to agriculture. Of that, 60% is wasted. Besides, several chemicals and pesticides are used excessively and often find their way into the water supply. Here’re some of the solutions that have emerged.
Solution: Predicting Water Usage by Individual Users in a Day #
A researcher at the University of Cordoba, Spain, developed a decision-making prediction model that predicts individuals’ water usage daily. It takes in various input variables, ranging from easily quantifiable ones such as temperature, humidity, and plot land to be watered, in addition to more complex values describing traditional watering methods in the area and holidays during the watering season. Then, by applying a specialised ‘fuzzy logic system’ model, curves are established for the inputs, and a neural network creates a relationship between them. Finally, it can predict the estimated millimetres of water usage. This data allows water user associations to accurately manage the water supply and prepare for maintenance, repair, and supply issues in advance without wasting water or negatively impacting areas requiring it.
Solution: Helping New Farmers With Water Usage #
In Japan, a government-supported digital farming solution involves collecting data from soil and light sensors. Then, it uses artificial intelligence for computers to advise farmers on the quantity of water and fertilisers used. These data-driven insights are invaluable to new, inexperienced farmers who may otherwise be guilty of wasteful water usage and allow them to increase their productivity.
Solution: Studying Water Usage in Fields to Find Issues and Improve Usage #
ConserWater, founded by Caltech graduate Aadith Moorthy, is a cost-effective A.I. monitoring system that tracks water distribution data in a field through satellites and historical data with no need for ground sensors or manual inspection. Then, it uses artificial intelligence to give recommendations for fine tuning the irrigation supply and watering process to attain optimal soil moisture and simultaneously minimising excess water consumption. It also allows users to identify leaks in irrigation pipes. Users can access the system through a mobile app or desktop website.
Solution: Using Artificial Intelligence to Allow Cotton Farmers to Use Pesticide Effectively #
Over 55% of India’s pesticide is used in cotton farming, but overuse of it often damages crops and quality. That’s why the Indian research institute Wadhwani AI built an artificial intelligence model to determine how many pests are in the area and send notifications to farmers advising them on pest usage. Over 18500 farmers use the low-resource intensive mobile app to send images of pests in the region and receive advice from the app on how much pesticide to use using three levels of alerts — green, yellow, and red. Experiments showed that this led to a 25% improvement in cotton crop yields due to more effectively managing pests and not wasting resources.
Solution: Only Using Chemicals to Eliminate Invasive Plants Where Needed #
Weeds that compete with neighbouring crops for resources such as sunlight, water, and nutrients are known to cost the farming industry tens of billions of dollars a year, which is why many farmers use chemicals generously to combat them. Farmers often have to spray chemical products all around the field to eliminate them, which leads to much waste and inefficiency. That’s why a team of researchers at the French Institut National des Sciences Appliquée developed a model that can take images captured by drones of regions of a farm (the model is currently restricted to fields of beets, spinach, and bean). Then, it labels areas with heavier weed concentrations. This allows farmers to focus on specific areas at the right time to use chemicals and reduce waste efficiently.
Problem: A Shrinking and Aging Labour Force #
Even as the worldwide demand for agricultural output increases, the number of farmers to produce it is shrinking in number and reducing in age. Artificial intelligence-powered innovations are therefore playing a vital role in filling this industry gap.
Solution: Harvesting Crops and Picking Produce Using AI-powered Robots #
Traditionally, produce on agricultural land is picked by farmers physically. But as their strength in numbers reduces, agrarian robots are taking their place, which are capable of bulk harvesting at a much-increased accuracy and speed without requiring vacation or time off. These machines help improve the yield’s size and reduce waste from crops being left in the field. There are several variations of this category of robots. Harvest Automation’s HV-100 is a farmer’s assistant that can move potted plants with ease and precision. Several cost-effective fruit picking systems also exist — whether it’s citruses with the Energid Citrus Picking System that can pick citruses every three seconds, the Agrobot E-Series that in addition to picking strawberries can identify individual strawberry’s maturity, a robotic vacuum-powered apple picker, an EU campaign-backed robot called Sweeper that picks peppers based off their colour, or an MIT robot gardener that connects wirelessly to several sensors attached to plants which call the water robot.
Solution: Using Driverless Tractors #
In addition to using AI-powered physical solutions, self-driving tractors are also an effective method of combating agricultural labour force issues. Driverless tractors with several advanced AI and non-AI powered features can ease some of the burdens on farmers. Tractors can autosteer themselves in a straight line and change direction without any farmer intervention allowed due to a combination of AI and GPS technology. Tractors such as Mahindra & Mahindra’s driverless tractor can steer to the next row for continuous operation without any intervention of the driver.This allows them to spray, plant, plow and weed cropland without any human intervention necessary.
Problem: Climatic and weather-related events #
As we saw in the last article, climate and weather-related events can ravage farmers. Over ten years between 2004 and 2014, agriculture lost 100 billion USD to such events.
Solution: Using A Smartphone App To Help Farmers Stay Profitable Despite Adverse Impacts of climate change #
PlantVillage is a firm that provides a free app to farmers in Africa to help protect their stable crops even in the face of global warming and changing weather patterns due to climate change. The app combines a variety of data, from the UN’s WaPOR (Water Productivity through Open access of Remotely sensed derived data) portal, weather forecasts, African soil datasets, and the United Nations Crop Calendar to provide critical information such as drought tolerance of crops and which crops are suitable in which areas. Furthermore, farmers can learn about climate-resilient crop varieties, affordable irrigation methods, flood mitigation, and soil conservation strategies, and other ideal practices to help them adapt to changing conditions. 
Solution: Using Satellite Data and AI to Predict Weather and Crop Sustainability #
Poor weather can devastate farmers of any stature, and climate change-induced erratic weather patterns have further accentuated that issue. A Colorado-based company called aWhere attempts to address it by providing detailed and custom weather data for individual farmers, crop consultants, and researchers. The company claims that it allows for access to over a billion points of agronomic data daily, including temperature, precipitation, wind speed, and solar radiation, and historical comparisons to anywhere else in the world. 
Problem: Pest Control and Crop Disease #
As we’ve seen previously globally, pests and crop disease reduce the output of five major food crops by anywhere from 10 to 40 percent and pose a significant threat to the global food supply, environment, and human health. Here’s how the agricultural sector is using AI to combat them.
Solution: Using Machine Learning to Predict Pest Behaviour Weeks in Advance #
Spain’s agricultural ministry has developed a technology that leverages artificial intelligence and big data to allow olive farmers to anticipate and prepare for the olive fly pest. A machine learning model analyses data collected by the Andalusian Plant Protection and Information Network (RAIF) on the pests and other crop-related factors to predict their behaviour and subsequently predict the percentage of olives likely to be eaten by the fly. This allows for superior decision-making, focusing more on areas and times of increased risk and impact, and taking actions to control the effect more sustainably and efficiently. 
Solution: Using a Smartphone App to Diagnose and Solve Issues Regarding Pests and Diseases #
Plantix, an idea that began with farming tribes and the demand for pest and disease problems, provides farmers information about pests and diseases using machine learning and scientific image data provided by the Indian-based ICRISAT (International Crops Research Institute for the Semi-Arid Tropics). By taking photos of affected crops using the free smartphone app, farmers can have pest damage, 500 different kinds of plant diseases, and nutrient deficiencies diagnosed using artificial intelligence with eighty-five percent accuracy. The app also provides a forum for users to discuss possible problems and solutions with each other or paid experts. In that manner, farmers can minimise the amount of crop and output lost to pests and disease. 
Problem: Financial Issues and Burdens #
As we saw in our first article, several farmers worldwide face critical financial issues — procuring financing for near-necessary purchases such as farm inputs and harvesting crop, significant debt, and lack of financial records and data.
Solution: Using AI to Predict Yield Forecast to Optimise Financial Aspects #
An API developed by IBM leverages big data and machine learning to provide personalised insights to farmers into the yield for corn crops up to three months in advance with limited computing resources. This data can determine yields for past growing seasons, vital for financial aspects such as validation of agriculture insurance claims and risk. Also, this allows farmers to optimise supply-and-demand chain logistics and predict commodity prices.