CHOICES
A publication of AAEA
Artificial intelligence (AI) is beginning to change how farming works in the United States. Most discussions focus on automation, meaning machines replacing workers. That is important, but it is only part of the story. An equally important and less discussed shift is augmentation, where technology helps people do their jobs better instead of replacing them.
That distinction matters because agriculture is not a typical workplace. Farm work depends on weather, biology, local knowledge, and timing. It also depends on judgment. Machines can collect data; human judgment remains essential for interpretation and action. That is why augmentation may be especially important in agriculture and why it deserves much more attention than it has received so far (Villacis, 2026).
In practice, this means that the role of the farmer is starting to shift. Instead of doing every task manually, farmers are increasingly managing systems, interpreting data, and making decisions with the support of AI tools—whether through digital extension agents, software platforms, or automated equipment. But this shift will not affect everyone equally. Without support for training, infrastructure, and data security, many farmers may struggle to keep up. That process is already underway.
AI is already being used in agriculture in systems that scan crops, guide machinery, monitor animals, and help managers make decisions. But this shift did not start with AI. Farms have been moving toward automated and data-driven systems for decades. For example, US milk production from farms using computerized milking systems increased from 20% in 2000 to 45% in 2021, and milk production from farms using computerized feeding systems increased from 22% in 2000 to 52% in 2021 (See Figure 1). These changes reflect a broader transition toward system-managed production, where technology plays a larger role in day-to-day operations.
More recent tools build on this foundation by incorporating AI directly into farm decisions. A clear illustration is John Deere’s See & Spray, which utilizes AI algorithms and cameras to identify weeds and pests, yielding considerable gains compared to traditional broadcasting and reducing herbicide use by 58% (Avent et al., 2026). Technologies like this do not just automate tasks; they also change how those tasks are carried out.
As a result, what farmers do on a daily basis is already changing. Instead of directly milking cattle or manually adjusting nozzle pressure, farm operators increasingly focus on maintaining sensors, interpreting software outputs, and overseeing automated systems. But the next phase of AI goes beyond physical systems.
Earlier technologies in agriculture focused on automating physical tasks. The newest wave of technology works differently: It involves systems that can read documents, answer questions, and generate useful information on demand. These tools, often referred to as generative AI, are designed to support cognitive tasks—the kinds of activities that involve processing information, generating responses, and supporting decisions—rather than replace manual labor. Recent work by Eloundou et al. (2024) shows that large language models (LLMs) such as ChatGPT expose many tasks to AI. They suggest that approximately 80% of US employees face at least 10% task exposure, while around 19% of workers may see over 50% of their tasks affected. However, exposure does not tell us whether a job will be replaced or improved.
In agriculture, this difference is especially important. Farm operators face a wide range of cognitive challenges that go beyond physical labor. AI assistants may help them navigate government regulations, manage contracts with suppliers and off-takers, or provide technical guidance (Shaikh et al., 2025). Some
of these tasks may be automated, while others may be improved by AI support. Many will change in ways that are not captured by a simple replace-or-not question.
This is where generative AI could have a particularly broad impact. By lowering the cost of accessing technical knowledge, these tools can make information that was once specialized more widely available. For smaller operations in particular, the growing availability of AI-based guidance could represent a meaningful shift in how decisions are made and who has access to expertise. To understand what this means for agriculture, it helps to distinguish between two ways AI changes work.
AI is changing agriculture through both automation and augmentation, and the difference between the two is important. Automation replaces labor in routine tasks (Restrepo, 2024; Villacis, 2026). Augmentation, by contrast, changes how people work by improving how they gather information, interpret it, and respond to it. Understanding this distinction helps explain how AI reshapes jobs.
A broader shift in the labor market helps put this into perspective. Over time, repetitive routine work has declined, while nonrepetitive, judgment-driven work has expanded (see Figure 2). Consider an accountant: In the 1980s, they might have spent 60% of their time ondocument processing and 40% on complex judgment calls. In the 2020s, with AI and automation, that split might look more like 20% document processing and 80% complex judgment calls. This suggests that technological change does not simply replace labor; it also increases the value of tasks that depend on judgment, interpretation, and adaptation.
This distinction is especially relevant in agriculture. Some farm tasks are relatively easy to standardize. Milking, spraying, planting, and harvesting can often be carried out by machines or guided systems. Other tasks are far less predictable. Crop stress, animal health, weather shocks, labor shortages, and equipment failures all require judgment. These are the situations where AI is more likely to support a worker than replace one. That is why augmentation is such a promising concept for agriculture.In many cases, this means AI is often used alongside human decision-making rather than instead of it. A farmer may use AI to detect disease earlier but still decide how to respond. A livestock manager may rely on sensors to track animal health but interpret those signals using experience and context. A farm operator may receive detailed summaries of supplier contracts but negotiate the final terms. In each case, AI alters how the human role is carried out.
This is where the policy conversation becomes more focused on adoption and support. AI is already affecting agriculture, and the key question now is how it will be used and whether it will strengthen both productivity and human capability. This shift is already visible in how farm work is changing.
The biggest change from AI may be the reorganization of farm work. As AI spreads, farmers are more likely to spend time managing systems, reading data, and making decisions based on software output. This raises the value of technical skills, problem solving, and oversight.
One way to understand this shift is to recognize that farmers do not perform a single job—they perform many tasks. Farm operators supervise workers, manage equipment, interpret field conditions, and handle business administration. Each of these responsibilities includes both routine activities and judgment-driven decisions. This helps explain how AI affects work. Using a task-based framework, we can consider how different activities are exposed to AI (Brynjolfsson, Mitchell, and Rock, 2018). Some tasks—such as scheduling, coordinating, or estimating—are well suited to AI support because they rely on processing structured information. Others, such as choosing agricultural techniques, selling products, or monitoring animal behavior, depend more heavily on experience and context and are less easily shaped by AI. Even where AI is well suited to a task, it does not necessarily replace the human role. In practice, these activities still require judgment, local knowledge,and adaptation to changing conditions. AI can generate useful analytics or recommendations, but human interpretation remains essential.
Evidence from the broader labor market points in the same direction. Over time, there has been a shift away from repetitive work toward nonrepetitive, judgment-driven activities. Jobs in this category rose from 29.5% of US employment in 1980 to 43.0% in 2020 (See Figure 2). Although detailed data for agriculture remains limited, descriptive evidence suggests a similar pattern. Drawing on six focus group discussions about precision agriculture, Ogunyiola, Stock, and Gardezi (2025) report that participants described farming as “becoming more mental than physical labor” and argued that “the best farmers are observational data collectors.” One farmer also noted that they “envision managing a fleet of robots.”
Another way to see the scope of this change is through the range of tasks AI may influence. The effect of AI on the labor market is sometimes described as an iceberg. On the surface, AI appears to affect a relatively small share of jobs—roughly 2.2% of the US economy. But beneath the surface, a much wider range of activities in administration, financial services, and professional services is exposed. Including these indirect effects, AI could impact up to 11.2% of the US economy (Chopra et al., 2025). A similar dynamic is likely in agriculture, where AI may not replace entire jobs but can reshape many of the tasks within them. These patterns suggest that farm work is being reorganized. Some routine tasks may decline, while others that rely on interpretation, adaptation, and oversight become more important. The result is a shift toward a form of work that is less about physical execution and more about managing systems and making decisions.
The way farm work is changing also depends on who is able to adopt new technologies. Historically, larger farms have been more likely to adopt advanced tools than smaller operations. This pattern has already appeared in earlier waves of precision agriculture (McFadden, Njuki, and Griffin, 2023; Lim et al., 2024). Data from the USDA Economic Research Service (USDA ERS) show a substantial gap in adoption. Only 13% of small family farms have adopted yield monitors, yield maps, or soil maps, compared to 68% of large family farms (Lim et al., 2024). Similar differences appear across other technologies, including guidance autosteering and variable rate systems. For example, McFadden, Njuki, and Griffin (2023) find that corn farms in the top quintile of acreage are more than five times as likely to adopt variable-rate technologies, such as computer-assisted herbicide application (see Figure 3). These patterns suggest that access to capital, scale, and technical capacity plays a major role in determining who benefits from new tools.
The next wave of AI-powered technologies may reinforce these differences. Some innovations, particularly those involving specialized hardware, arelikely to remain expensive and capital intensive. Strawberry harvesting—a notoriously labor-intensive process—may soon rely on commercially viable autonomous picking robots (Tituaña et al., 2024; He et al., 2025). As with earlier precision agriculture tools, these systems are likely to be adopted first by larger operations.
The uneven adoption of agricultural technology raises an important question: Can some forms of AI lower these barriers? Compared to expensive AI-powered machinery, generative AI tools are, at present, relatively low cost for end users. This creates the possibility that not all technological advances will follow the same path as capital-intensive equipment.
In some cases, this shift is already beginning. Tools such as the Utah PeachBot, developed by researchers at Utah State University Extension (Hill and Narine, 2026), and ExtBot, a free extension service model developed by the Extension Foundation (Extension Foundation, 2026), are designed to provide farmers with accessible, AI-based guidance. Unlike specialized machinery, these tools do not require large upfront investments, making them feasible for a wider range of operations.
In principle, this type of technology could expand access to information and technical support. Farmers who may not have the resources to adopt advanced equipment could still benefit from decision support, recommendations, and real-time guidance. While this does not eliminate the structural differences in adoption, it suggests that generative AI may play a role in narrowing some of the gaps.
Differences in technology adoption point directly to the role of policy in the age of AI. If AI is changing farm work in important ways, then the challenge is adaptation. Farmers need support to understand when AI tools are useful and how to use them effectively. Workers need opportunities to build skills that match a more technical and data-driven agricultural system. Farms also need access to digital infrastructure and data protections that allow them to participate in this transition on fair terms.
Although US agriculture does not yet have a single, comprehensive policy framework for AI, responses to these challenges are already beginning to emerge. Extension systems are adapting to the digital landscape. For example, North Carolina State University has introduced “AI 101” workshops, and partnerships with nonprofits such as the Extension Foundation have supported the development of agriculture-focused generative AI tools like ExtBot. These efforts point to one clear policy direction: strengthening extension services and public training programs so that farmers are not left to navigate AI on their own.
At the same time, infrastructure remains critical. Broadband expansion supported through federal policy continues to play a central role, as many AI tools depend on reliable high-speed internet. Without it, small and medium-sized farms may struggle to access even low-cost digital technologies. In addition, data governance is becoming increasingly important. Initiatives such as Ag Data Transparent (2024) highlight the need to ensure that farmers retain control over their data. As in other parts of the digital economy, who owns the data will shape who benefits from these technologies. What happens next will depend on how these choices are made.
There is also a broader lesson about how technology affects work. Automation and augmentation are not opposites. They often move together. Acemoglu and Restrepo (2020) show that technology can both replace labor and create new demand for complementary tasks. Recent work by Acemoglu, Kong, and Restrepo (2025) makes a similar point by emphasizing how technology changes tasks within jobs. That is exactly what is happening in agriculture. AI is replacing some work and reshaping the tasks that remain.
This points to a future of farm work shaped by ongoing change. Some jobs may shrink, others may grow, and many will be reorganized. The larger risk is that AI’s benefits will be unevenly distributed. Farms and workers that are already well positioned may gain the most, while others struggle to adapt.
That is why the policy challenge is to support both automation and augmentation. This means helping farmers adopt useful tools, supporting workers as tasks evolve, and ensuring that smaller operations are not left behind. In agriculture, augmentation is not a side effect of AI—it may be one of the most important ways the technology changes the sector. If policy supports this transition well, AI can make agriculture more productive while still relying on human judgment, local knowledge, and better forms of work. If it does not, the gap between farms that can adapt and those that cannot may continue to grow.
Acemoglu, D., and P. Restrepo. 2020. “Robots and Jobs: Evidence from US Labor Markets.” Journal of Political Economy 128(6): 2188–2244. https://doi.org/10.1086/705716
Acemoglu, D., F. Kong, and P. Restrepo. 2025. “Tasks at Work: Comparative Advantage, Technology and Labor Demand.” Handbook of Labor Economics 6:1–114. https://doi.org/10.1016/bs.heslab.2025.08.003
Ag Data Transparent. 2024. “Core Principles.” Available online: https://www.agdatatransparent.com/principles
Avent, T.H., J.K. Norsworthy, T.R. Butts, G.L. Drescher, and L.L. Nalley. 2026. “Targeted Herbicide Cost Requirements for Adoption of See & Spray Technology and Ways to Improve Area Sprayed.” Agrosystems, Geosciences & Environment 9:e70324. https://doi.org/10.1002/agg2.70324
Brynjolfsson, E., T. Mitchell, and D. Rock. 2018. “What Can Machines Learn, and What Does It Mean for Occupations and the Economy?” AEA Papers and Proceedings 108:43–47. https://doi.org/10.1257/pandp.20181019
Chopra, A., S. Bhattacharya, D. A. Salvador, A. Paul, T. Wright, A. Garg, et al. 2025. “The Iceberg Index: Measuring Skills-Centered Exposure in the AI Economy.” arXiv 2510.25137. https://doi.org/10.48550/arXiv.2510.25137
Eloundou, T., S. Manning, P. Mishkin, and D. Rock. 2024. “GPTs Are GPTs: Labor Market Impact Potential of LLMs.” Science 384(6702):1306–1308.
Extension Foundation. 2026. ExtensionBot. Available online: https://extension.org/tools/extbot/
Figliola, P.M., B.E. Humphreys, C.L. Pechtol, and L. Zhu. 2021. The Infrastructure Investment and Jobs Act (P.L. 117-58): Summary of the Broadband Provisions in Division F. Congressional Research Service Report R46967. Available online: https://www.congress.gov/crs-product/R46967
Gillespie, J., E. Njuki, and A. Terán. 2024. Structure, Costs, and Technology Used on U.S. Dairy Farms. USDA Economic Research Service, Economic Research Report ERR-334.
Gleason, T. 2025, November 5. “Transforming Farming Through AI.” Closing Market Report. Illinois Extension. Available online: https://extension.illinois.edu/podcasts/closing-market-report/special-transforming-farming-through-ai
He, Z., Z. Liu, Z. Zhou, M. Karkee, and Q. Zhang. 2025. “Improving Picking Efficiency Under Occlusion: Design, Development, and Field Evaluation of an Innovative Robotic Strawberry Harvester.” Computers and Electronics in Agriculture 237(C):110684. https://doi.org/10.1016/j.compag.2025.110684
Hill, P.A., and L.K. Narine. 2026. “Large Language Models for Agricultural and Rural Development: An Application of Foundational Models in Extension.” Advancements in Agricultural Development 7(2):23–34. https://doi.org/10.37433/aad.v7i2.603
Lim, K., J. McFadden, N. Miller, and K. Lacy. 2024. America’s Farms and Ranches at a Glance: 2024 Edition. USDA Economic Research Service, Economic Information Bulletin EIB-283.
McFadden, J., and K. Lim. 2024. “Precision Agriculture Use Increases with Farm Size and Varies Widely by Technology.” Charts of Note. USDA Economic Research Service. Available online: https://www.ers.usda.gov/data-products/charts-of-note/chart-detail?chartId=110550
McFadden, J., E. Njuki, and T. Griffin. 2023. Precision Agriculture in the Digital Era: Recent Adoption on U.S. Farms. USDA Economic Research Service, Economic Information Bulletin EIB-248. https://www.ers.usda.gov/publications/pub-details?pubid=105893
North Carolina State University Plant Sciences Initiative. 2025. “Extension Agents Explore AI for Agriculture.” Available online: https://cals.ncsu.edu/psi/news/extension-agents-explore-ai-for-agriculture/
Ogunyiola, A., R. Stock, and M. Gardezi. 2025. “Precision Agriculture and the Future of Agrarian Labor in the US Food System.” Agriculture and Human Values 42:383–403. https://doi.org/10.1007/s10460-024-10615-x
Restrepo, P. 2024. “Automation: Theory, Evidence, and Outlook.” Annual Review of Economics 16(1):1–25. https://doi.org/10.1146/annurev-economics-090523-113355
Shaikh, T.A., T. Rasool, K. Veningston, and S.M. Yaseen. 2025. “The Role of Large Language Models in Agriculture: Harvesting the Future with LLM Intelligence.” Progress in Artificial Intelligence 14:117–164. https://doi.org/10.1007/s13748-024-00359-4
Supic, N., and K. Josifidis. 2025. “Will Generative AI Bring Change: Technological Disruption and Redistribution in the United States?” Journal of Economic Issues 59(2):392–399. https://doi.org/10.1080/00213624.2025.2493528
Tituaña, L., A. Gholami, Z. He, Y. Xu, M. Karkee, and R. Ehsani. 2024. “A Small Autonomous Field Robot for Strawberry Harvesting.” Smart Agricultural Technology 8:100454. https://doi.org/10.1016/j.atech.2024.100454
US Department of Agriculture, National Agricultural Statistics Service (USDA-NASS). 2025. Farm Labor. Available online: https://esmis.nal.usda.gov/publication/farm-labor
US Government Accountability Office. 2024. Precision Agriculture: Benefits and Challenges for Technology Adoption and Use. GAO-24-105962.
Villacis, A. H. 2026. “Artificial Intelligence (AI) and Agribusiness: From Automation to Augmentation in a Global Context.” Agribusiness forthcoming. https://doi.org/10.1002/agr.70062
