Haleigh Reichen | October 20, 2026
https://www.deltaplexnews.com/local-news/from-irrigation-to-robotics-uapb-research-examines-ais-growing-role-in-agriculture/
Artificial intelligence may be best known for chatbots and image generators, but researchers say some of its most practical applications could be taking place far from a computer screen – in fields, greenhouses and other agricultural operations.
A recently published article authored by University of Arkansas at Pine Bluff researchers, led by Dr. Shahidul Islam, professor, graduate coordinator and director of UAPB’s USDA Regulatory Science Center, examines how artificial intelligence is used to monitor crops, detect diseases and pests, manage irrigation, predict yields, and even guide agricultural robots.
The study, “Artificial intelligence in modern agriculture: recent advances, challenges and future directions,” reviewed 95 peer-reviewed studies published from 2021 through 2025. It found rapid growth in agricultural uses of artificial intelligence, while also identifying significant obstacles that continue to limit widespread adoption.
“Our study found that agricultural AI has increasingly moved beyond individual applications toward systems that combine sensors, computer models and automated equipment to make decisions in real time,” he said.
For farmers, that could mean technology capable of determining when a crop needs water, identifying signs of disease before they become widespread or using information about soil and weather conditions to help determine how a crop should be managed.
One of the most promising uses identified in the study is precision agriculture, Dr. Islam said.
“Instead of treating an entire field uniformly, precision agriculture uses information about specific conditions to help producers make more targeted decisions,” he said. “AI systems can analyze data on soil properties, weather, crop health and satellite imagery to inform decisions about irrigation, fertilization, pest management and harvesting.”
Smart irrigation provides a practical example.
Sensors can continuously measure factors such as soil moisture, temperature and weather conditions. Machine-learning systems can then analyze that information to determine when plants need water and how much should be applied.
“We reviewed studies in which AI-supported irrigation systems reduced water use or improved water-use efficiency while maintaining crop health,” Dr. Islam said. “The technology could become increasingly important as farmers contend with water scarcity and changing environmental conditions.”
AI is also being used to identify crop diseases and pests. Cameras mounted on equipment, drones or other systems can collect images of plants, while computer models learn to recognize visual signs associated with weeds, insects, disease or plant stress.
“In some cases, identifying a problem earlier could allow producers to respond more precisely rather than applying treatments broadly across an entire crop,” he said.
The technology is also changing controlled-environment agriculture, including greenhouses, hydroponic systems and vertical farms.
Modern greenhouse systems can use networks of sensors to monitor temperature, humidity, light, irrigation and nutrients, Dr. Islam said. When those sensors are combined with AI, the system can analyze changing conditions and automatically adjust parts of the growing environment.
“Our review found that environmental monitoring was the most frequently represented application among the studies examined,” he said. “Greenhouse and protected cultivation, along with pest and disease detection, were also major areas of research.”
Robotics represents another rapidly developing area.
Agricultural robots are being developed for tasks including harvesting fruit, grading mushrooms, monitoring crops and performing other repetitive or labor-intensive work. Researchers found that robotics has begun moving from laboratory concepts toward practical agricultural tasks.
“AI can also help machines perform jobs that are difficult to automate,” Dr. Islam said. “A harvesting robot, for example, may need to identify a piece of fruit, determine its position among leaves and branches and decide how to grasp it without causing damage.”
Despite the rapid progress, Dr. Islam and his fellow researchers caution that agricultural AI is not ready to solve every problem on the farm.
Many of the systems examined in the review remain at the prototype or proof-of-concept stage. Relatively few have been extensively tested across different farms, crops, seasons and environmental conditions. As a result, a system that performs well in a laboratory or controlled greenhouse may not necessarily produce the same results under changing field conditions.
“Cost is another barrier,” Dr. Islam said. “Sensors, cameras, drones, automated equipment, internet connectivity and computing infrastructure can require substantial investment. Smaller farms and operations in areas with limited digital infrastructure may have greater difficulty adopting the technology.”
The researchers also identified shortages in technical skills, limited standardized agricultural datasets and concerns about whether AI models developed for one crop or environment will work reliably in another.
“Data ownership, privacy and cybersecurity present additional questions as farms collect greater amounts of information through sensors, drones and cloud-based platforms,” Dr. Islam said. “Our study shows that clear policies on ownership, consent and data sharing will be important for building farmer trust.”
Those limitations mean that the future of agricultural AI will depend on more than developing increasingly sophisticated computer models.
“More real-world testing is needed, along with lower-cost and more accessible systems, improved training and Extension services, and technology designed around farmers’ needs,” Dr. Islam said.
Researchers are also placing greater emphasis on “explainable AI,” which is designed to help users understand how an AI system reached a recommendation rather than simply providing an answer.
“That transparency is particularly important in agriculture because farmers need to be able to understand and trust the information they are using to make decisions,” he said.
The study also documents how quickly agricultural AI has evolved. From 2021 to 2022, much of the research centered on sensors, crop monitoring and yield prediction. By 2025, researchers were increasingly studying interconnected sensor systems, digital models of farming operations, autonomous robotics and AI systems designed to make their decisions more understandable.
“This represents a movement toward what researchers call ‘Agriculture 4.0,’ in which sensors, artificial intelligence, automation and farm equipment increasingly work together,” Dr. Islam said.
For producers, however, the significance of that transition may be much simpler: having better information about what is happening on the farm and being able to respond more precisely.
“AI has substantial potential to support more productive, resource-efficient and climate-resilient agriculture,” Dr. Islam said. “But for these technologies to be widely adopted, they need to be more than accurate in a research setting. They also need to be affordable, reliable and practical under real farming conditions.”