Article  | 

The role of Artificial Intelligence and Machine Learning in modern farming


Reading Time: 4 minutes

The traditional image of the farms that produce our food is familiar to us all. The farmhouse, the large wooden barns, and the golden fields of wheat under a blue sky. An old-fashioned idyll in which the most modern technology to intrude on the vision is the combine harvester. This oldest of industries might not bring Artificial Intelligence, Machine Learning and cutting-edge tech innovation to mind.


But that image is increasingly separated from the modern reality of farming. The fact is that agriculture is facing a crisis, and only technology can save it. Farming is going to fall behind the rising population curve, and will struggle to feed every mouth, in the face of an increasingly urbanised world and a sharply rising population.


Currently, 38 percent of the world’s available land, and 85 percent of the available fresh water supply, go into agriculture. And these figures are rising. The old practices are unsustainable, and food waste is only exacerbating the problem. But AI and new tech is already working to disrupt the industry, and offer solutions that could guarantee a more sustainable food supply for all of us into the future.


Tech disruption from seed to table


AI and new tech is making a difference in almost every aspect of modern farming. Even before seeds are planted, data on how certain seed types respond to given soil conditions, weather and diseases are facilitating crop selection and hybridisation. Once those seeds germinate, image recognition software is learning to differentiate crop sprouts from weeds, with a successful identification rate of over 75 percent and rising. Those weeds no longer need to be removed by hand, either; laser weeding, using computer-controlled, tightly focused infra-red light to disrupt cell growth, is already consigning that labour-intensive activity to the past.


Automation already goes beyond weeding, though. Many people will at least have seen video footage of large machines crawling across fields, expertly removing vegetables from the soil. These robots are already much more efficient than traditional harvesting by hand. But robotic agricultural technology has broadened into almost every role where human labour is expensive or scarce, or resource consumption is inefficient.


Smart irrigation, for example, uses automated drip irrigators which are connected and controlled wirelessly. Artificial Intelligence is used to analyse automatically gathered data on humidity, wind speed, solar radiation, crop density and soil properties. This data is used to reduce water use and increase yield simultaneously. And drone technology keeps a watchful eye over the whole process. Monitoring crop health, drones have already proved themselves invaluable by providing much more accurate imagery than was previously available. Drones provide up-close images which are then processed using image recognition software and AI in combination, reporting on crop maturity or alerting the farmer when diseases are spotted.


Cutting out the supply chain


This revolution is already happening. A 2020 study estimated that a typical modern farm will be processing 4.1 million data points per day by 2050. But for some farmers, the entire structure of our food supply system is wrong. They’re combining new tech with AI to completely change how and where food is produced in the first place. In a recent interview with Harvard Business Review, Mike Zelkind of 80 Acres Farms explained that the supply chain would be better removed completely.


Mike’s company wants to use smaller, vertical farms and cutting-edge technology to both dramatically speed up crop cycles and produce food more sustainably. Mike’s firm is leveraging Machine Learning and AI to scale the operation from hobby levels to a commercial enterprise. They’re disrupting the model that’s become the norm, where food is increasingly produced for longevity and durability rather than nutrition. With so many intermediate handlers between farm and table, perishable produce is now being produced with the aim of lasting 30 days so it survives the supply chain in saleable condition. This has a considerable detrimental effect on taste and nutrition.


Vertical farming uses smaller, indoor agriculture, in locations within the communities that are going to consume the food. Locals are recruited where possible, and food miles are reduced to almost zero. The blend of tech and locality is producing quite startling advances in sustainable production. Using just a tiny fraction of the space of a traditional farm, and 95 percent less water, crop cycles vary from 17 to 28 days, depending on the produce. This compares against months-long crop cycles of existing farming techniques.


The greater yields are down to what’s been learned from the wide variety of data available. Machine Learning is used to analyse the effects of both ambient and root-medium temperature fluctuations. Air flow, micro-climate conditions, crop health and yield curves are all monitored by AI, in addition to the plants’ responses to the stresses induced on them deliberately to speed up maturation. Mike describes the arrangement as a marriage between ML, data analytics and robotics on the one hand, and agri-science on the other.


There are, however, some criticisms levelled at hydroponic or vertical farms compared to larger-scale methods such as greenhouse production. One is that the energy used for LED lighting makes it less efficient than simple sunlight. But modern LED lighting, and smart monitoring and analysis of light levels and wavelengths, make the losses from seed to table considerably lower than losses incurred in the standard supply chain. Meanwhile, energy comes from renewable resources, so the energy consumed for lighting is sustainable, in line with the overall model.


Ultimately, advances in farming technology could mean that the days when a lettuce was chilled, sprayed with chemicals and carried across an entire continent to be sold 30 days after it was harvested seem laughable. While vertical and hydroponic farming are not going to take over from the larger-scale operations any time soon, the application of AI, Machine Learning and new tech is quickly becoming essential to food security, regardless of scale or method.


The Collider is a venture-building programme that brings together researchers, corporates and entrepreneurs to enhance our ecosystem and create a positive impact on our society as a whole. Powered by Mobile World Capital Barcelona, the 2021 edition focuses on the use of Agriculture 4.0.