Tianlong Chen believes in precision agriculture

For his CoreAI research showcase presentation earlier this year, Tianlong Chen tapped his agricultural roots: His parents have a blueberry farm in southwest China.

“The daytime is very hot. The nighttime is very cold,” the ideal combination for creating sugar in fruit, said the assistant professor in the UNC College of Arts and Sciences’ computer science department. “That is a perfect place you can have large and very sweet blueberries.”

Chen was thinking of farmers like his parents in creating “Agent and Robot Teaming for Precision Agriculture,” a project using drones to scan fields for diseased plants and robots to treat them.

The project is a rare example of artificial intelligence in agriculture and an example of the “human-centric” approach of Chen, winner of Electronics magazine’s 2025 Young Investigator Award.

“We focus on human-centric AI applications,” he told the magazine. “We prioritize AI applications with a broader societal impact, particularly within the biomedical sector, to enhance the quality of daily life for the general public.”

Chen’s research focuses on building accurate, trustworthy and efficient machine learning systems. At Carolina, he helped create a socially responsible chatbot to provide reproductive health information and salvaged old graphics processing units to use them to train AI models.

Chen also wants to use AI to improve the lives of farmers here in North Carolina, a state where agriculture generates more than $100 billion in economic impact annually. Many farmers already use automation to plant, fertilize and harvest their crops and rely on computers for information about weather and crops. Chen’s goal is to make automation more intelligent by integrating it with information.

Chen combines AI and robotics to improve farming in North Carolina — using real data to build smarter systems that support farmers in the field. (ITS Communications/Kate Chapman)

“Tons of reasons motivated me to do this,” Chen said of the project, a collaboration with Lirong Xiang, a former assistant professor at NC State University who is now at Cornell University. “My imagination is that we will ultimately be integrating the systems across human, AI and the robots.”

In the project, the farmer could tell the AI agent the disease symptoms spotted and even upload photos. Then the AI agent could analyze the data to identify the disease and recommend treatment. The farmer then reviews the agent’s response and triggers the treatment using a robot, trained to move to certain coordinates through something like a computer game.

In a presentation video, Chen showed a spider robot being deployed into a small research plot of tomato plants, moving across and down the rows to remove a diseased leaf or to apply a treatment.

This kind of precision is still at the research stage, Chen said, but AI is already helping farmers determine the right density of plants for their acreage. At the rate AI is developing, farmers may be able to use his research in, “if everything goes smooth, maybe a couple of years,” he said. By then, the farmer could probably use a phone instead of a computer, too, and the costs of the tech should decrease.

Still, precision agriculture like this won’t work for every farm, especially the smaller ones. “The keyword will be ‘scalability,’” Chen said, explaining how a robot can more efficiently do tedious work like regularly scanning hundreds of acres of crops.

“So AI actually didn’t replace the role for humans, but it serves as a tool to help humans to scale up and improve the productivity and improve efficiency,” he said. Using AI may also make it easier for a farmer to try new crops, offering more flexibility.

Robots aren’t always the answer, Chen admitted. His parents still harvest blueberries the old-fashioned way, by hand.

“How to do the harvest is one of the fundamental challenges for the robots,” he said. “The fruit is very easy to be destroyed. It’s very soft.”