[Expert Column] In the Era of Autonomous Agriculture, Safety Standards Must Evolve Together

[Expert Column] Lee Kwang-wook, Head of Daedong’s Agricultural AX Promotion TF Division
 

Korean agriculture now stands at a major turning point. The aging of rural communities, chronic labor shortages, and growing production uncertainty caused by climate change show the reality that it is difficult to guarantee a sustainable agricultural future through conventional methods alone. Amid these changes, AI-based autonomous agriculture technology is drawing attention as a core technology for the future of farming.

In particular, the importance of autonomous operation technology is growing, enabling agricultural machines such as tractors, rice transplanters, and combines to recognize their surroundings, make decisions on their own, and carry out tasks. Advanced agricultural countries view autonomous agricultural machinery as an important means of securing future agricultural competitiveness and are accelerating technology development, while in Korea as well, AI-based autonomous agriculture technology is being developed rapidly.

Daedong is also preparing in earnest for the era of autonomous agriculture by launching Korea’s first AI tractor this year. Daedong’s AI tractor goes beyond the simple level of GPS-based straight-line assistance and is equipped with technology that can recognize and judge actual farmland environments on its own. The six vision cameras mounted on the vehicle recognize paddy and field boundaries, crop conditions, and obstacles in real time. Combined with RTK-GPS, a high-precision positioning system, and Daedong’s self-developed on-device AI technology, the agricultural machine is designed to determine work routes and perform tasks autonomously.

When people encounter this kind of technological progress, many naturally ask one question. "If tractors can now work on their own, doesn’t that mean people are no longer needed?"

Why It Remains at ‘Safe Automation,’ Not Full Unmanned Operation
To state the conclusion first, not yet. According to current domestic safety standards related to autonomous agriculture, agricultural machinery operating autonomously must be monitored by the user from nearby during operation. Daedong’s AI tractor is also designed so that autonomous operation is possible only when the user’s smartphone is within a certain distance of the tractor. This is not so much due to a lack of technology as it is due to institutional standards in the early stage of the industry that place top priority on safety.

Agricultural machinery moves in environments that are far more complex and unstructured than those of ordinary automobiles. There are various risk factors, including paddy ridges and slopes, sudden ground changes, hard-to-predict obstacles, and unexpected approaches by people or animals. When an accident occurs under fully unmanned conditions, many aspects of who bears responsibility—the user or the manufacturer—also remain legally and institutionally unclear.

For these reasons, most autonomous agriculture technologies at home and abroad currently remain at the stage of ‘safe automation’ rather than ‘full unmanned operation.’ Although autonomous driving technology is being actively developed in major agricultural markets including North America, it is difficult to say that fully unmanned work with no human intervention at all has become commonplace.

A New Safety Device: AI Vision Technology
However, the development of autonomous driving technology does not necessarily mean only new risks. Rather, as vision-based recognition and decision-making technology becomes more advanced, it is highly likely to develop in a direction that reduces accidents that can occur when people directly operate agricultural machinery.

Farmers can accumulate fatigue during long hours of work, and due to momentary carelessness or blind spots, they may recognize hazards on slopes and paddy ridges too late. In contrast, cameras and sensors mounted on agricultural machinery continuously observe even areas that are difficult for the driver’s line of sight to reach. 

As technology that analyzes ground inclination, surrounding obstacles, and vehicle movement in real time advances, it will be possible to detect rollover risks in advance, stop work, or change routes. Functions that recognize the possibility of collisions with people or other agricultural machinery early and slow down or stop will also become more sophisticated. In other words, in the future, AI and vision cameras will constantly monitor the work environment and go beyond replacing human judgment to become new safety devices that compensate for risks people may miss.

Establishing Institutional Safety Performance Standards in Step with Technological Evolution
Recently, AI vision technology, obstacle detection sensors, remote monitoring systems, and precision position control technology have been advancing rapidly. For now, users must monitor work conditions from nearby, but if technology becomes sufficiently advanced and accident-prevention performance is verified, safety standards will also need to evolve accordingly.

For example, beyond deciding whether to allow autonomous operation based on whether the user is near the agricultural machine, it is also possible to consider establishing standards centered on actual safety performance—such as how accurately the machine can recognize dangerous situations and stop safely, and how effectively it can prevent rollovers and collisions.

What we should pay attention to is the fact that the goal of autonomous agriculture is not simply to create ‘agriculture without people.’ The true goal is to reduce the labor burden on farmers and create a new agricultural system that can carry out farm work more safely and efficiently.
 

Technological Development and Institutional Improvement for a Sustainable Agricultural Future
The automobile industry also was not capable of full autonomous driving from the beginning. Along with technological development, safety standards, laws and institutions, and social consensus have evolved step by step. The same is true for autonomous agriculture. Now, in line with the pace of AI technology development, safety standards and related systems must evolve together.

For Korean agriculture to move into a true era of autonomous agriculture, not only companies but also the government, industry, and research institutions must work together to find a balance between safety and technological advancement. If technology can provide higher safety than existing methods, a system must also be established to objectively verify this and reflect it in institutions.

The future of agriculture does not lie simply in unmanned work without people. Only when human experience and AI’s judgment capabilities complement each other, and technological innovation and safety standards develop together, will Korean agriculture finally gain safer and more sustainable competitiveness.