Although artificial intelligence–based agricultural tools are developing rapidly, one of the main obstacles in Hungarian farming is not the algorithms themselves but farmers’ readiness to use them, said Varga Péter Miklós, vice-president of Magyar Precíziós Gazdálkodási Egyesület, in an interview with Szegeder. He argued that a large segment of Hungarian agriculture struggles to adapt to digital solutions.
An example is the crop-monitoring system developed by the University of Szeged (Szegedi Tudományegyetem). The AI-based tool can detect plant stress before visible symptoms appear, allowing farmers to be alerted to water or nutrient shortages early and intervene before yield losses occur.
The issue is use, not hardware
Varga emphasized that such developments are important steps to help agriculture adapt to climate change and volatile markets. Yet, he believes that adoption will depend far more on users’ digital skills and openness than on the precision of the AI.
He pointed to the electronic farm logbook (egn) as an example: producers must record plant protection treatments, nutrient applications and other agrotechnical operations digitally, but many still find data entry problematic. “We are talking about seven data points in total that must be recorded after a spray,” Varga noted, adding that if such a relatively simple administrative task is difficult for many, it is unclear how quickly more complex, AI-based decision-support systems will spread.
For this reason, Varga argued, the solution is not more standalone apps but systems that integrate into interfaces farmers already use. If an interface could provide immediate recommendations or alerts after data entry, it would likely be used more consistently.
Not only large farms stand to benefit
Digital adoption is also slowed by farmers’ tendency to trust personal experience or peer advice more than data-driven decisions. Nonetheless, Varga said, more producers are gradually opening up to precision technologies, and AI-based solutions need not be exclusive to large operations.
The sensors, external databases and services that form the basis of such systems typically cost on the order of tens to hundreds of thousands of forints. Building a full precision-farming system can require much larger investments, potentially in the hundreds of millions of forints, but a decision-support solution based on AI can be useful with a far smaller outlay. Therefore, a smaller producer can still see a return if an AI system helps detect a disease or problem early that would otherwise cause significant yield loss.
Varga said he sees no strict constraints tied to farm size, crop or sub-sector: data collected by small farms can be valuable because an early local detection of disease or pest can inform other producers. Thus, digital data collection can benefit not only an individual farm but the broader sector.
Long-term data collection may be the real competitive edge
According to Varga, the ultimate measure of success for developments like the SZTE system will not be AI accuracy alone but whether they provide practical help in everyday farming. That requires sufficient quantities of high-quality data. He therefore advises farmers to record as much information about their fields as possible now, even if they do not yet see exactly how those data will be used.
It has long been standard in agriculture to preserve all available information. Notes on weather, soil condition or crop development may seem sparse today, but within a few years AI models could be built on those records to draw useful conclusions. Farmers who start collecting digital data only years later will miss historical series that cannot be retroactively reconstructed, leaving little to build an accurate system on; the longer the time series, the more precise the AI forecasts.
Varga therefore believes that data collection itself could become at least as important a competitive advantage as the technology. AI does not create value on its own but through the information that farmers consistently record over many years.
Education and professional oversight remain necessary
While AI can support more and more agricultural tasks, Varga stressed that it is still far from replacing farmers’ decisions. Algorithms can analyze data and make recommendations, but interpreting and validating those suggestions remains the user’s responsibility. He compared this to medicine: just as doctors had to learn to use new diagnostic tools, farmers must understand how AI-based systems operate so they can judge what data underpin each recommendation and how reliable it is.
“AI should also be taught and examined,” he said, suggesting that AI-based services should be subject to ongoing professional validation similar to other agricultural professional systems.
Varga sees plant-monitoring and precision-farming solutions becoming part of everyday agricultural toolkits over the next decade. He noted that some “hands-free” operations already exist where algorithms perform tasks and the resulting data continuously improve system performance. In farms that consistently collect data over long periods, it may eventually be possible for AI to make certain decisions with fewer errors than humans — but that requires years of reliable data and continuous professional oversight.
Photos: Bálint András / Szegeder
The author is Farkas Anna, a reporter for Szegeder; this article appears on Telex as part of a cooperation between Szegeder and Telex.
