Rice Fields Enter the Era of Artificial Intelligence
One day, the scarecrow may become nothing more than a relic of the past. Amidst expanses of ripening rice, it will no longer be straw dolls standing guard, but intelligent cameras, artificial intelligence (AI) algorithms, and automated devices capable of recognising pests, diagnosing plant diseases, and even deterring birds with precision without damaging the environment. This vision is no longer science fiction. An agricultural technology revolution is underway, and Indonesia has a significant opportunity to become an important part of it.
This momentum comes at a very opportune time. Food security has now become one of the most strategic issues facing the nation. Global climate uncertainty, geopolitical conflicts, disruptions to the world food supply chain, and a continuously growing population mean that the ability to produce one’s own food is no longer just an option, but a necessity.
Indonesia has placed food self-sufficiency as one of its national development priority agendas. This target certainly deserves appreciation. However, self-sufficiency will never be achieved merely by expanding planting areas or increasing the cropping index. The biggest challenge actually lies within the rice fields themselves, when farmers must face pest attacks that are increasingly difficult to predict.
Data provides a fairly clear picture of this situation. Up to August 2024, Indonesia’s rice imports increased by approximately 121.34 per cent compared to the same period the previous year. This figure indicates that national productivity is still under considerable pressure. On the other hand, climate change is causing pest and plant disease attack patterns to become more dynamic. Unpredictable seasons, rising temperatures, and changing humidity create an ideal environment for the development of various plant pests.
For decades, farmers have relied on field experience to determine when to carry out control measures. This method proved successful when seasonal patterns were relatively stable. However, current conditions have changed. Climate change has caused pest behaviour to change as well. Many species are developing faster, expanding their distribution areas, and even appearing at times that were previously never anticipated.
In such a situation, speed becomes a decisive factor. Pests can no longer be waited for until their populations are clearly visible. Once symptoms of an attack appear, economic losses have actually already begun. Therefore, the ability to detect early is the main key in modern agricultural systems.
This is where artificial intelligence offers a very fundamental paradigm shift.
The development of AI in agriculture is no longer just about helping to process data, but has become a decision support system that works in real time. Development projections published in the Safarinda Journal for the 2025–2026 period show an increase in the accuracy of AI-based detection systems from around 82 per cent in 2023 to around 96 per cent in 2026. In the same period, the incidence of pest attacks is projected to increase from around 38 per cent to 50 per cent.
These two figures convey an interesting message. Nature is indeed changing so that pest pressure is increasing. However, technological capabilities are developing faster than the adaptation rate of plant pests. With an accuracy rate approaching 96 per cent, AI systems can recognise early symptoms before an attack develops into a population explosion that is difficult to control.
More importantly, this technology changes the philosophy of pest control. So far, success has often been equated with the amount of pesticide used. In fact, this approach has created various new problems, ranging from pest resistance, the death of natural enemies, environmental pollution, to increased production costs.
Modern agriculture is moving in a different direction. What is needed is not more pesticides, but more precise decision-making.
One interesting example comes from an innovation developed by researchers at the Kalimantan Institute of Technology through the Automatic Turret Gun, an AI-based automatic bird deterrent system. The name of the technology may sound like a military device, but its function is entirely aimed at crop conservation.
The Javan munia (Lonchura leucogastroides) and the white-headed munia (Lonchura maja) are two species that can cause crop losses of up to 30–50 per cent, with potential economic losses of around Rp2.5 million per hectare. For years, farmers have relied on scarecrows, plastic ribbons, or noises to deter them. Unfortunately, birds have a high learning ability, so these methods quickly lose their effectiveness.
The Automatic Turret Gun offers a much more adaptive approach. By utilising the YOLOv8 algorithm, the system is able to recognise and track bird movements in real time with an mAP@50 value reaching 97.6 per cent. A camera moves to follow the target via a servo motor, while gel pellets are used to deter the birds without injuring the animals or leaving chemical residues.
Innovations like this demonstrate that technology does not have to conflict with the principles of sustainable agriculture. On the contrary, artificial intelligence enables control to be carried out more effectively while also being more environmentally friendly.
A similar transformation is also occurring in plant disease diagnosis systems. Previously, the application of AI was often considered only possible if a fast internet connection was available. This assumption is beginning to be debunked through edge computing approaches, where data processing is carried out directly on location without having to send all information to a cloud-based computing centre.