KANSAS CITY, MISSOURI, US — The history of agriculture in general, and grain production in particular, can be described as a never-ending pursuit of efficiency in the quest to improve outcomes to feed more people. In this regard, technology has always played a pivotal, almost relentless, role in creating tools, whether used by a singular farmer or large-scale grain handling and storage operations, to meet the moment.

For advocates of artificial intelligence (AI), the moment has arrived for AI and the many tools it promises to bring to bear on the world’s production of staple cereal crops and oilseeds. Like any tool, AI’s usefulness may be seen in the problems it solves or the processes it makes more efficient. 

Downtime in the grain industry is costly. Improving reliability and enhancing safety with predictive maintenance supports operations and profitability. Accurate measurement of material flow and weights optimizes processes, reduces waste and promotes productivity.

The goal of AI would be to work smarter by augmenting human insight and experiences with enhanced decision-making capabilities, according to industry observers. AI is growing exponentially in complexity, and facilities must have the right infrastructure in place and the right knowledge and support to take full advantage. Responsible implementation is deemed critical, as a strategy that does not align with existing applications and company objectives can be counterproductive and create data security issues.

Temperature monitoring, level monitoring, ambient temperature monitoring, moisture monitoring, fan control, access to bin data, etc., are all areas where AI could prove invaluable to production and profitability. Sustainability and stewardship of the environment are being demanded by governments and consumers, which requires greater data collection and interpretation than previously has been necessary.

“The pace of adoption of AI technology in grain handling and storage varies across different organizations and regions,” said Zachary Carr, president of SafeGrain, a grain quality control and hazard monitoring company based in Dayton, Ohio, US. “While some early adopters are leveraging AI solutions to gain a competitive edge, others are more cautious due to factors such as cost, infrastructure limitations, and the need for specialized expertise. However, as awareness grows and AI becomes more accessible, I anticipate a steady increase in adoption rates, especially as the benefits become more evident.”

CESCO EPC GmbH is an internationally operating German company based in Konstanz that designs and supplies industrial plants for grain logistics and deep processing with handling, storing and milling systems. Martino Celeghini, chief executive officer of CESCO, is similarly enthusiastic about AI’s potential. 

“The grain storage industry is on the cusp of an intelligent revolution,” Celeghini said. “Artificial intelligence is rapidly transforming how facilities operate, optimizing processes, maximizing efficiency and safeguarding grain quality. From predictive maintenance to automated quality assessment, AI is poised to significantly enhance every step of the grain storage cycle.”

Dirk Maier is a professor of agricultural and biosystems engineering at Iowa State University in Ames, Iowa, US, a longtime educator and renowned expert in grain storage and handling with extensive experience in developing public-private funded university-based training facilities. Maier is director of the recently inaugurated Iowa State University Kent Feed Mill and Grain Science Complex, which utilizes state-of-the-art technology in its operations. While acknowledging AI’s potential to further drive improvements in grain storage and handling, Maier said today’s advanced technology has set the stage.

“We have the kind of sensing technologies, monitoring technologies that continue to improve and become more readily accessible, and people will readily adopt them to keep track of the quality and the predictability itself,” Maier said. “Models can basically run online, interpreting data being collected in a storage bin in terms of temperature, moisture, and in particular, CO2, and predictively tell you about trends that could lead to a hot spot in your bin. Those are already things that are available and will become even more available as the computing technology improves.”

The surging interest in AI in grain handling and storage comes as little surprise to Vikram Adve, professor in the Computer Science Department at the University of Illinois at Urbana-Champaign, Illinois, US. Adve co-founded the Center for Digital Agriculture and leads the AIFARMS Institute, one of the national AI research institutes launched under the aegis of the US National AI R&D Strategic Plan: 2019 Update.

“There are various products that you can buy and use that do rely on AI, but I would say the degree of adoption is relatively limited relative to its potential, and in fact agriculture as a domain, broadly speaking, is significantly behind quite a few other domains in terms of adoptions of AI, which just means there’s a lot of low-hanging fruit,” he said. “There’s a lot of capabilities that can make a significant impact, and there’s really a large market potential right now for new products and new technologies to be adopted.”

What is AI?

AI itself might provide a unique, almost human answer to the question of its origin, even today as the technology continues to evolve.

The recent attention given to AI in the public consciousness likely can be traced to the 2022 emergence of OpenAI’s ChatGPT, an AI system that uses a large-language model to engage users in humanlike conversation and churn out topical written documents, which seemed to jump-start global excitement and questions about its potential.

The concept of AI goes back several decades to the earliest computers, when programmers pondered the very idea that a machine could replicate human thought patterns and intuitively learn new skills or generate unique ideas. As the processing power of computers grew, so did expectations for AI technology.

However, ChatGPT and other language models are but one subset for AI.

Adve described AI as a particular class of techniques, with some aspects that have been extremely successful and some that have not been very successful at all.

“AI is really a broad class of (computing) techniques, all of which broadly speaking aim to mimic or exceed human capability and their various aspects or functions in the world,” Adve said. “What aspects of AI are the most successful, the most widely used, have changed over time. Over the last 30 years, the most successful areas of AI have evolved quite dramatically, starting with small successes in handwriting recognition, speech understanding and language, dramatic advances in computer vision, some significant advances in robotics and autonomous systems, and most recently, in the past five years, dramatic advances in language understanding and language generation.”

Adve said chatbots using large-language models, such as ChatGPT, generate a lot of excitement due to the way they can interact with humans and be used by people on a day-to-day basis, which has not been true for most previous generations of AI. Other aspects of machine learning and computer vision could perhaps prove even more valuable to agriculture over the long term.

“Humans still far outstrip AI systems and are likely to do so for a long time when you think of the totality of what we can do, but there are aspects of scale, for example, that AI systemscan significantly outperform humans by learning patterns from enormous amounts of data,” he said.

Celeghini said a transition is underway from traditional programmed algorithms to neural network-based systems of AI to enhance grain handling and storage operations.

“In this paradigm, behavior is primarily shaped by training data rather than explicit programming, enabling handling of multiple inputs simultaneously and generating complex responses that are difficult to achieve through conventional programming languages,” he said. 

Carr sees the same potential for AI to mimic human intelligence to perform tasks such as learning, problem solving and decision making.

“Unlike traditional predictive systems that rely on pre-programmed rules and algorithms, AI systems can analyze large datasets, detect patterns, and make autonomous decisions without explicit programming,” Carr said. “Essentially, AI enhances predictive capabilities by continuously learning from data, adapting to new information, and improving its performance over time.”

AI on the job

Analysts generally agree AI is no replacement for humans, who ultimately will continue to make the decisions when it comes to grain storage and handling. But as a tool, AI does hold great promise for all the systems that currently help feed the world’s 8 billion people by ensuring that crops taken from the field are optimally stored and protected.

AI-powered solutions increasingly are being integrated into commercial elevators and farm storage operations, Celeghini said. These solutions are utilized in various aspects, such as predictive maintenance to anticipate equipment failures, optimizing grain storage conditions by monitoring temperature and moisture levels, and enhancing grain quality management through automated sorting and quality control processes.

“AI can be particularly useful in optimizing operational efficiency and enhancing decision-making processes in grain storage and handling,” he said.

For instance, AI algorithms can analyze historical data to predict market trends, optimize inventory management, and minimize waste. Additionally, AI-powered monitoring systems can detect anomalies in grain storage conditions in real-time, enabling proactive interventions to prevent spoilage and ensure grain quality.”

In addition to quality and hazards sensors, Celeghini sees optimization of space as particularly impactful for efficient operations.

“AI algorithms can analyze years of historical data on grain inflow and outflow to create intelligent storage allocation plans,” he said. “This ensures that every cubic meter of storage space is utilized effectively, eliminating inefficiencies and minimizing the energy wasted on maintaining empty silos and ensuring first-in/first-out approach.”

Carr said AI holds tremendous potential to transform various aspects of the grain storage and handling industry. Some potential areas where he said AI could contribute include predictive maintenance to minimize equipment downtime and reduce maintenance costs; optimization of inventory management to prevent overstocking or shortages; real-time monitoring of grain storage conditions to prevent spoilage and ensure quality; automated grading and sorting of grains based on quality parameters; data-driven decision-making to optimize logistics, pricing, and market strategies.

“AI has the power to revolutionize how grain storage and handling operations are managed, leading to greater efficiency, sustainability and profitability,” Carr said.

In an article Maier wrote for World Grain in 2019, he posited several prescient questions about technology for corporate and cooperative grain handling companies that apply as well to AI: How rapidly can they prepare for innovative technologies to remain competitive? How many employees will be needed to operate their facilities going forward? What knowledge and skills will future employees need? What continuing education and training will employees need?

“If I’m a cooperative with 45 locations, how can I optimize my system?” Maier said. “Generative AI in the future will help make those types of strategic decisions, hopefully more objectively or at least provide recommendations that are objective, and boards and executives will have to then make decisions based on those recommendations. And they can run a lot more what-ifs in there as well.”

Where Maier sees AI potentially useful in the most human sense is as an interactive repository of knowledge gained by longtime employees that could be accessed by newer workers with questions about how to approach a problem. Perhaps AI could even interpret all that gained knowledge and personal experience in new ways that become more useful.

“In the end, we’re still producing food, fuel and fiber that is going to come from a field that’s going to have to be aggregated someplace, that will have to be processed in whatever form it needs to be for utilization,” Maier said. “These are supply chain physical processes that we, hopefully, can do in the future smarter and more efficiently, but the physical processes still have to happen.”

Adve said advances in generative AI already are proving their worth throughout the agriculture value chain, but particularly at the farm level where advisory services traditionally obtained from an extension office can be provided at a fraction of the cost; visual data that can be analyzed by AI to provide solutions for pests or disease in the field; and data-driven decision making by generative AI that will become a lot more sophisticated in the years ahead. Many of these advances will be especially helpful in less-developed regions that may lack reliable support infrastructure.

“Even if a tool is capable of doing an automatic answer, I think it’s much more effective if there’s a human in the loop making decisions on how to take advantage of that or how to apply it,” Adve said. “I think we add to human capabilities by incorporating generative AI tools and many other AI tools as well, and that is really by far the most effective way to take advantage of this.”

What’s next?

As with any technology, Carr said there are several factors to consider when AI enters the conversation on a farm or at a commercial elevator.

“Firstly, they should assess their current operational challenges and determine whether AI solutions can address these effectively,” he said. “Secondly, they should evaluate the cost-benefit analysis, considering factors such as potential cost savings, improved efficiency, and competitive advantages. Thirdly, they should assess their readiness in terms of infrastructure, workforce capabilities, and organizational culture to embrace technological change. Ultimately, the decision to adopt AI should be driven by a strategic vision for long-term growth and sustainability.”

CESCO’s Celeghini said there’s a general interest among customers, with some companies embracing AI as a marketing tool. However, full integration into existing systems and equipment is just getting started.

“One hurdle is the complexity of AI implementation in settings like grain storage plants, which typically rely on straightforward automation,” he said. “Customized training for each project adds cost that may not always justify the benefits. Instead, focusing on specific AI applications like online quality analysis or predictive maintenance could yield more feasible solutions, independent of plant specifics.”

Maier said it was the construction of the Kent Feed Mill and Grain Science Complex at Iowa State that provided a refreshing perspective on technology. The importance of people with specialized skills who can build a facility and implement new technologies such as AI will not diminish, he said.

“We need people with those types of hands-on, specialized technical skills, and that, too, will continue in our plant when it comes to maintenance,” he said. “People probably will need to have a higher knowledge because of the types of systems that our plants will use in terms of automation. You’ve got to understand not just the wiring, but also how all that relates to the PLC (programmable logical controller) and how all of it ties into the automation system. That’s not going to get replaced by AI at all.”

Adve said adoption of AI in agriculture could be slowed by a lack of expertise. Ag companies understand the problem and understand their customers but may not have the AI proficiency in-house to develop AI-based products and services. To address this shortage, the Center for Digital Agriculture, in collaboration with the University of Illinois’ highly respected departments of Agriculture and Biological Engineering, Crop Sciences and Computer Science has developed the first asynchronous online Master of Engineering in Digital Agriculture degree and professional certificat programs.

“We think it can be valuable in expanding the workforce with kids coming out of school today and the existing workforce, because there is a vast workforce in agriculture and ag tech companies that would benefit from this training,” Adve said. “They already have strong ag experience and understanding of the markets and the customer needs, so giving the training they need in a couple of AI topics would be able to start working on products that could have a greater impact in the near term.”

SafeGrain’s Carr emphasized the importance of collaboration and knowledge sharing within the industry to harness the full potential of AI and other emerging technologies. Addressing concerns about cybersecurity and data privacy is paramount, especially given the sensitive nature of operational data. Additionally, the lack of sufficient technology infrastructure in rural or less-developed areas may pose challenges in terms of connectivity and access to AI-enabled solutions. Moreover, there may be resistance to change and a need for workforce training to ensure seamless integration and utilization of AI technologies, he said.