| Will AI Help Agriculture Flourish? AI has already had a large impact on agricultural productivity and recent leaps in the technology will help sustain that momentum. |  |
| In just a few short years, artificial intelligence has evolved from a technological curiosity to an almost singular obsession in the public consciousness. With trillion-dollar companies racing to apply the technology to every area of human endeavour, the rest of us are left to ask what it will mean for our careers, our industries, and our societies. Clearly, AI has enormous potential and will spur transformational change. This note attempts to cast light on one sliver of the issue with some high-level thoughts for the agriculture sector.As a starting point, it’s worth recognizing that AI is nothing new, nor is its application to agriculture. It was in the 1950s that researchers began thinking about how to train early computers for tasks that require humanlike intelligence, as opposed to just straightforward calculations. Progress was slow, but over the decades growing computing power, advances in methodology, and the accumulation of vast amounts of training data produced steadily better results. By the early 2010s, AI systems were being used to drive recommendation engines, recognize objects in photos, and in a host of other applications. What’s new in the 2020s is that AI can now understand and generate language, video and other types of highly complex material—and it can do it increasingly well.In the agriculture sector, researchers began experimenting with AI as far back as the 1980s. Early systems could provide crop management guidance based on environmental inputs, but a lack of available data limited their use. As elsewhere, adoption only started to become widespread in the 2010s, motivated by the need to extract actionable information from satellite imagery and a growing constellation of sensors monitoring fields and livestock. Today, AI-driven systems are relied upon widely to make highly granular decisions about crop and livestock management. Actions once taken for an entire field or barn are now being optimized for individual animals and plants. |
| All of this, coupled with improvements in plant and animal genetics, has propelled a long boom in farm productivity. Statistics Canada estimates that multifactor productivity, which measures the amount of output produced from a given quantity of inputs, increased 41% in agriculture and adjacent industries between 2000 and 2023 (Chart 1). That is by far the strongest performance of all Canadian industries and contrasts with a slight decline for the business sector as a whole. It would be wrong to attribute these gains entirely to AI, but the technology has greatly amplified the impact of better data collection by helping convert billions of disparate data points into usable information for farmers and their equipment. |  |
| The relatively early roll-out of AI in agriculture means that some of the low-hanging fruit has already been picked. On farms, the most obvious uses for the technology do not involve the types of complex content that it recently became capable of handling. For computers, using drone data to inform a crop sprayer is a lot easier than writing a poem about it. And now, since agricultural machinery doesn’t last much more than a decade, there is less pre-AI legacy equipment being phased out. New systems are still more productive, but incrementally so. Partly because of this, Canadian farm productivity leveled off during the first few years of the 2020s, though the numbers were also dragged down by poor growing conditions. In the United States, where data are available through 2024, farm productivity has continued to grow.Setting aside fluctuations in the weather, farm output should continue to benefit from more expansive data collection and better AI systems to make sense of it. For example, newer systems can use real-time optics to enable sprayers to target individual weeds, even once a crop has emerged. AI is also driving advances in tactile equipment, which can physically test ripeness and harvest delicate products like mushrooms, which have until now been cultivated by hand. In the livestock space, emerging technologies can automatically dispense optimized diets for individual animals, informed by real-time behavioural and physiological monitoring. These technologies would mostly have been possible without recent advances in AI, but they will further boost productivity and should yield ongoing benefits as they self-learn over time.There is also potential for more radical new innovations. The latest advances in AI could push its contribution far beyond monitoring and responding to conditions on the farm. Perhaps most promising is the possibility of improved drug discovery (for both humans and livestock) and faster progress on plant and animal genetics. Over time, AI systems could also improve weather and climate forecasts, which would allow farmers to respond more proactively to conditions on the ground. Future systems could also take a more holistic view of the farm as a business, helping to guide big-picture strategic decisions and resource allocation. |
| Rising farm productivity should also help keep a lid on food prices. Canadian grocery prices have jumped more than 30% in the past five years, far outpacing the 20% increase in the overall consumer basket and creating a major pain point for consumers. To be fair, grocery prices are driven by more than just agricultural costs, so rising farm output will not be a panacea. Over the past half century, food prices in both Canada and the United States have increased roughly twice as much as farm prices, as products have grown to include more processor-added value than the simple ingredients of generations past. This has naturally made non-agricultural costs more important at the grocery store, but farm-level developments still exert a significant influence on consumer food prices (Chart 2). At the very least, rising farm productivity should help keep food prices closer in line with overall inflation, especially in an environment of slower global population and food demand growth. |  |
| As with any new technology, the AI roll-out will also create challenges. A big one for farmers is that same tendency for rising productivity to weigh on agricultural prices. Although consumers may benefit, the overall impact on farmers will depend on the cost of AI technologies, the increase in crop and animal production, the resulting impact on selling prices, and the savings achieved on inputs like crop treatments and feed. Another complication is that few producers will want to fall behind on technology, even if the immediate economics are not ideal. However, any challenges in the agriculture sector will be small compared to those in the information and knowledge industries, where the AI roll-out will be much more profoundly transformative.Bottom Line: AI has already had a large impact on agricultural productivity and recent leaps in the technology will help sustain that momentum. There will also be challenges, but an industry that has managed centuries of transformation should be able to rise to the occasion. |