The AI/ML Revolution in Agriculture

As one of the biggest producers of agricultural goods in the world, the United States features domestic economic and social development that is deeply intertwined with the continual growth of its agricultural sector. In 2019, the American agricultural industry added $1.109 trillion to U.S. Gross Domestic Product (GDP), accounting for 5.2% of U.S. GDP. In addition, 13% of U.S. citizens’ household budgets are spent on food and agricultural commodities, while agricultural industries employ a total of 10.9% percent of employed people in the United States (1). These facts illustrate the importance of health, emphasizing the importance of the growth of the U.S. agricultural sector on American economic well-being in terms of contributing to GDP, pulling American spending back into the American economy, and by further enhancing the approximately 22.2 million jobs that the agricultural industry provides.

However, despite the importance and power of the agricultural sector in the United States, both new problems and missed opportunities within the industry have prevented agriculture from fully benefiting from potential opportunities for growth and better societal outcomes. Additionally, ecological challenges due to climate change pose a significant threat to crop production. Given the growth of the U.S. and global population, ecological changes that decrease crop production call into question the ability for the agricultural industry to meet these demands. Farmers face new criticism for the extent of their carbon emissions, negatively impacting societal outcomes, transit inefficiencies in distributing their goods, plant disease posed by climate change, crop production inefficiencies, and lack of uptake in farm mechanization. In parallel to the ecological challenges, it has been predicted that the total food production globally will need to increase by 70% from its existing levels in order to adequately feed the growing population (2). Therefore, although the United States has somewhat integrated digital platforms to increase agricultural production and better ecological outcomes that better society, the government should accelerate the integration of artificial intelligence (AI) and machine learning (ML) platforms utilized in other countries in order to evolve agricultural systems to better function with changing conditions. This integration, in turn, will allow the U.S. agricultural industries to pivot with changes in industry quickly, meet the growing demand of food production, and retain and increase the positive impact that the agricultural industry has on U.S. production.

The combination of AI/ML into farming techniques and within the agricultural industry as a whole offers potential for increased profits within the industry, increased job opportunities, new means to feed a growing population, and as a method to tackle the aforementioned ecological problems. More specifically, AI and ML techniques offer new opportunities for species management. When used in combination with the development of high computing power and the growing importance of big data, the use of AI and ML in agriculture offers methods to enhance the quality of species breeding by allowing for the analysis of “crop performance in various climates to build a probability model” to determine which plants will fare best in certain, changing climates. In addition, AI and ML can assist farmers in better managing their fields by creating models that analyze evaporation rates, soil temperature, and water content to minimize costs farmers spend maintaining their lands and instead maximize their crop production as ecological conditions fluctuate. Furthermore, farmers can use AI and ML techniques to better predict crop yield, crop quality, and the presence of weeds in order to more effectively price and allocate food grown in a way that benefits both farmers and consumers. AI and ML techniques can also predictively monitor animal welfare by tracking dietary needs and stress levels that assist in optimizing animal well-being while better understanding animal yield. Lastly, the use of AI and ML in agriculture can help increase the efficiency of food distribution logistics in such a way that minimizes cost and delivers food more efficiently.

In particular, the value of integrating AI/ML into agricultural production is especially highlighted by the development of technology such as Plantix, created by the German agricultural tech startup PEAT. The technology is able to monitor crop and soil health through deep learning techniques that can “identify natural defects and nutrient deficiencies” in crop soil. Plantix utilizes image recognition from smartphone cameras alongside algorithms to analyze these images to ultimately “correlate particular foliage patterns with certain soil defects, plant pests, and diseases” (3). After identifying abnormalities that Plantix is able to detect through photos, the platform identifies “soil restoration techniques, tips, and other possible solutions” to help the user address their crop and soil health in order to optimize crop production and maximize profits with the increased health, quality, and productive capacity of their soil and crops (4). Given that the United States Department of Agriculture has estimated that the agricultural industry loses approximately 44 billion USD annually due to soil erosion, efficient diagnostic technologies that streamline the process of detecting soil and crop health could greatly cut back on this loss, aid in food production, and offer solutions that are more favorable ecologically. Plantix already possess 500,000 users and is expected to continue to gain traction from farmers in the future.

Furthermore, companies such as Blue River Technology have also illuminated the power of AI and ML when utilized in the agricultural industry. In recent years, Blue River Technology has developed their See and Spray and LettuceBot technologies to help farmers more efficiently allocate herbicide use to the plants that actually need them. It has been estimated in the past that annually in the U.S., farmers utilize approximately 310 pounds of herbicide on their fields to mitigate weed growth, which is incredibly costly given that herbicide is traditionally sprayed on the entire field as it is challenging to distinguish what plants are weeds and which are not (5). In order to cut costs and provide a more ecologically sustainable solution to this problem, Blue River’s Lettuce Bot are towed behind traditional tractors spraying equipment with streamlined cameras as the tractors move across the crop fields (6). With these cameras, the LettuceBot is then able to utilize machine-learning to help determine which plants are weeds among the crops by examining 5,000 plants in one minute, and then targets specifically weeds with a pinpointed, automated sprayer. Ultimately, it is estimated that farmers could be able to reduce their herbicide usage by 90% with the use of the Lettuce bot, and already treats approximately 10% of the lettuce produced in the United States (7). After being acquired by John Deere in 2017 for $305 million, Blue River’s technology offers enormous upside to increase food production, cut costs for farmers, and is able to offer better ecological outcomes through this use of automation (8).

Given the value of integrating ML/AI techniques into agricultural production, there are still significant barriers that need to be addressed before these technologies are easily accessible and utilized by the United States agricultural industry. The development of these technologies require significant testing, development, investment, and production costs that make it challenging for agricultural producers to have access to these products early in their development. Additionally, there are significant educational and time barriers for farmers to not only acquire these technologies, but also implement them into their specific farms. In the future, cost of investment, the amount of time and research needed to make these technologies, and educational barriers will have to be addressed in order to enable AI/ML techniques to have the maximum impact in the agricultural industry. However, as exemplified in these cases, integrating AI/ML technology into agricultural production will not only help the agricultural industry to fully capitalize on its potential to increase economic growth and social production, but also help contribute to mitigate societies current most pressing problems such as problems resulting from global warming, labor shifts, and population growth.


Sources

  1. https://www.ers.usda.gov/data-products/ag-and-food-statistics-charting-the-essentials/ag-and-food-sectors-and-the-economy/
  2. https://www.syngenta.com/en/innovation-agriculture/challenges-modern-agriculture
  3. https://emerj.com/ai-sector-overviews/ai-agriculture-present-applications-impact/
  4. https://emerj.com/ai-sector-overviews/ai-agriculture-present-applications-impact/
  5. https://www.wired.com/2016/05/future-humanitys-food-supply-hands-ai/
  6. https://www.wired.com/story/why-john-deere-just-spent-dollar305-million-on-a-lettuce-farming-robot/
  7. https://www.wired.com/2016/05/future-humanitys-food-supply-hands-ai/
  8. https://www.wired.com/story/why-john-deere-just-spent-dollar305-million-on-a-lettuce-farming-robot/