Jakarta, INTI - The National Research and Innovation Agency (BRIN) continues to promote the use of artificial intelligence (AI) as a strategic technology to accelerate innovation in nuclear research. Through Politeknik Teknologi Nuklir Indonesia (Poltek Nuklir) of BRIN, AI is being utilized to support the development of nuclear materials, improve reactor operational safety, and even strengthen food research.
Frendy J. Kusuma, a researcher at Poltek Nuklir BRIN, explained that current scientific developments have entered the fourth paradigm, data-driven science, where large amounts of data are the primary foundation for generating scientific discoveries.
In this context, the application of machine learning is becoming increasingly important, particularly to accelerate research into nuclear materials, which have millions of possible configurations.
"For example, in the case of material optimization, there are millions of possible material configurations. The trial-and-error approach through conventional experiments or simulations is very costly and time-consuming. This is where machine learning is needed to find the best configuration more quickly and efficiently," explained Frendy on Tuesday, July 28, 2026.
AI Utilization for Nuclear Reactor System
According to Frendy, his postdoctoral research has implemented AI to develop a machine learning model capable of predicting the formation energy and band gap of actinides. Furthermore, AI has been utilized to predict the type and rate of two-phase flow in the natural circulation passive coolant system of a nuclear reactor, which plays a crucial role in supporting the design of a safer and more efficient reactor system.
The implementation of AI, Frendy said, supports the design and material development stages, as well as the operational phase of nuclear facilities, such as the Kartini Reactor.
"The development of AI in nuclear facility operations, such as early anomaly detection and predictive maintenance, has one fundamental end goal. To ensure operational efficiency and safety," said Frendy.
Innovation in Food Sector
For the food sector, Poltek Nuklir BRIN developed the application of machine learning to support quality control and radiation dose verification in the irradiation process of food samples, such as chicken, using a gamma irradiator. This approach is expected to improve the accuracy of the irradiation process while expanding opportunities for the development of nuclear-based food technology.
Frendy added that the success of AI development is determined not only by the ability to build models, but also by the quality of the data used. According to him, approximately 80% of the machine learning development process occurs in the data collection, data processing, and feature engineering stages. These stages provide a very promising area for generating novelty in research.
"In the machine learning development stage, the largest portion is not actually on model training. Around 80% of our work will be spent in the data collection, data processing, and feature engineering phases. However, it is precisely in these phases that the greatest opportunities for novel research can be explored," he concluded.
Conclusion
Poltek Nuklir BRIN utilizes AI to accelerate innovation in nuclear research, from material development and improving reactor operational safety to food research. Machine learning-based AI is used to predict nuclear material configurations, formation energies, actinide band gaps, and reactor coolant flow to support safer and more efficient designs. At the operational stage, AI is applied for early anomaly detection and predictive maintenance of reactor sensors, while in the food sector it is utilized for quality control and radiation dose verification.
Read more: Microsoft Expands Azure Platform AI Infrastructure with AMD’s Technology