Academic Impact
Cultivating AI in the Field of Agriculture
Artificial intelligence is no longer a future concept in agriculture. It is already shaping how food is grown, how data is analyzed and how the next generation of agricultural leaders is trained.
That balance between opportunity and responsibility shapes the scholarship of ALEC faculty members Dr. Chris Clemons, Dr. James Lindner, Dr. Jason McKibben and Dr. Jillian Ford, all of whom are deeply involved in teaching, researching and globally discussing artificial intelligence in agriculture. At the heart of it all, Clemons said, is the understanding that AI trained on massive datasets, known as Large Language Models (LLMs), must have careful human guidance.
“We have to remember that LLMs are stochastic (random) interpreters,” Clemons said. “Their entire training is on the probability of what word comes next based on the data provided. There is no native intelligence to it.”
In simple terms, LLMs do not “think” or “understand” the way humans do. They predict patterns — essentially guessing the next word based on massive amounts of data. Because of this, they can sometimes produce what appear to be confident but incorrect answers, a phenomenon known as “hallucinations.”
“While the term ‘Artificial Intelligence’ was coined back in 1956 at Dartmouth, we are still dealing with the constraints of the technology, specifically hallucinatory outputs that appear authentic but are factually incorrect. A hallucination exists because these models are stateless.”
Being “stateless” means the system does not truly remember past interactions the way a human would; instead, it reconstructs context each time, filling in gaps with probabilities. As a result, without the influence of human expertise, small errors can grow.
“If you are not managing this as the governor of information and the gatekeeper of fact, the potential for error is absolutely there,” Clemons said. “The human must remain the ‘Expert in the Loop.’”
Teaching AI as a tool, not a shortcut
That philosophy is embedded in ALEC’s new artificial intelligence course in agricultural education. Designed for both undergraduate and graduate students, the course goes beyond basic AI use and instead focuses on how the technology works, where it fails and how it can be responsibly adapted to real agricultural problems.
“The new AI course was born out of the consistent and historical need for agricultural education to always be at the forefront of technological adoption and implementation,” Clemons said. “We wanted to move beyond simply looking at AI at the surface level or teaching students just how to write a prompt.”
For Lindner, the course fits squarely within agriculture’s global mission.
“I hope that students take away that this is just another tool to help them contribute to solving the greater challenges facing agricultural systems today,” Lindner said. “We have an incredibly hungry planet with a population fast approaching 10 billion people.”
Those challenges — feeding more people with finite land and resources — are no longer “easy” problems to solve. Lindner emphasized that AI is not a replacement for expertise, but a way to enhance it. That includes understanding what AI can and cannot do.
As he explained, “We are not to the point where AI is making decisions for us. AI is not like ‘Skynet’ in the movies, where it has become self-thinking. We are just using vast computing power to solve problems that exist — for example, using AI tools to map a location to optimize the use of roller crimpers on a peanut crop to suppress weeds. We haven’t reached the point where generative AI is actually thinking for itself; we are using it to speed up the things we are already doing.”
AI in the field — and its limits
In agricultural practice, AI is already being used to process vast amounts of data that no human could reasonably analyze alone. From weather patterns and soil conditions to crop growth and livestock health, AI systems can rapidly detect trends and flag issues.
Still, McKibben cautioned against assuming anyone fully understands where AI is headed.
“I think the idea that we can answer that question ‘in general’ is a bit presumptuous,” McKibben said. “That would be like asking DARPA scientists in 1965 what the internet would be.”
Just as early internet researchers could not foresee today’s digital world, McKibben argued that AI’s final form remains unknown. What is clear, however, is the need for human capability to guide it.
“Without people driving it toward where we need it to be as a species, we will never reach its full potential; it will only ever be a ‘glorified Google,’” he said.
One reason ALEC faculty have leaned into AI adoption is the pace at which the technology evolves. Clemons described AI as improving at a rate that traditional research and education models struggle to match.
“Today’s model of AI that you are using is the most clunky, worst-designed, antiquated, slow AI model that you will ever use in your entire lifetime,” he said.
“I’ve heard this era defined by rapid obsolescence and described as ‘the age where AI will age like milk.’”
Failing to adopt AI now, they argue, risks leaving agricultural education and outreach behind.
From research to real-world tools
That urgency has translated into applied research and tool development. ALEC faculty are testing a ground-based AI program designed to help the Alabama Cooperative Extension System more quickly assess statewide program impact — allowing educators to evaluate what is working and where resources are needed.
Clemons described the work this way: “We look at that from the perspective of updating existing methodologies to meet the technological abilities we have now. Coming from the Ag Ed program, we have developed a few tools that are currently in testing. One tool we are really excited about is a ground-based program that will help the Alabama Cooperative Extension System better and more rapidly assess the impact they are having statewide. The intent is to allow agents to develop evaluations quickly to support their programming.”
This collaborative approach reflects a core ALEC philosophy. As Lindner noted, “In Agricultural Education, Leadership, and Communications, we are very fond of saying that we have ‘no pride in authorship’ and that we have to work together. The problems facing agriculture today are so significant and vast that no one of us can solve them. We are forced to work together, and we all bring unique tools to solve these problems.”
In addition to the Extension-focused tool, ALEC faculty are developing two patent-pending technologies. One, Veri-CAM™, addresses a common AI problem: fabricated citations. By verifying sources during the writing process, the tool helps researchers maintain academic accuracy and confidence.
Another project, a cooperative study with the University of Nebraska led by Clemons, Becky Haddad (University of Nebraska), McKibben and Lindner, explores AI-assisted photo elicitation — using AI to generate better prompts for images used in research and education.
“The core mission of the Alabama Cooperative Extension System is to provide scientific information to help people in the state make better decisions about their lives and agriculture,” Lindner said. “These tools allow us to design better and deliver educational programs that meet those specific needs.”
Global leadership on AI in agriculture
That work has also placed ALEC faculty on the international stage. In October, Lindner, Clemons, and McKibben attended and contributed to a global symposium on artificial intelligence, digital technologies and “Agriculture 4.0” at University College Dublin.
Hosted through the Journal of Advancements in Agricultural Development, the conference brought together scholars from around the world to share research on how AI is reshaping agriculture. Twelve papers from the symposium were published in January 2026, with McKibben and Clemons presenting as leaders in artificial intelligence within agricultural education and Lindner serving as a journal and symposium organizer.
Looking ahead, Clemons described a key challenge in AI adoption known as the “Black Box Paradox” — systems that produce answers without showing how they arrived there. ALEC’s work is focused on shifting toward “Glass Box” systems, where data sources and decision pathways are transparent.
By building specialized, problem-focused AI agents, rather than relying on one-size-fits-all models, researchers can improve reliability.
“What we can do is build our own Agentic Systems, focused agents that act as highly transformable, highly interpretive ingesters of knowledge focused on a specific problem,” Clemons said. “By doing that, you essentially remove all of the extra ‘noise’ from the model. The real power is not the generalized model, but the ability to ingest focused data for a specific response. It is the difference between a General Practitioner and a Specialist.”
And despite AI’s growing role, one thing is clear:
“We are not to the point where AI is making decisions for us,” Lindner said. “We are just using vast computing power to solve problems that exist.”
For ALEC, the future of AI in agriculture is not about replacing people, but about equipping students, educators and producers with better tools — and the knowledge to use them wisely.