Trends & Viral Yonhap News Culture · 2026-11-09

South Korea Unveils Agricultural AI Trained on 36,000 Scientific Papers to Revolutionize Farming Research

Key Points
  • South Korea's Rural Development Administration has developed a specialized agri-biotech large language model trained on 36,000 scientific papers and research reports.
  • The model provides verified answers with direct scholarly citations and is scheduled to open to the public through the NABIC platform starting in January.
  • The project is part of a broader roadmap to build a collaborative AI by 2028 capable of formulating scientific hypotheses and interpreting complex agricultural big data alongside human researchers.
Summary South Korea's Rural Development Administration (RDA) has taken a significant leap into the future of smart agriculture by unveiling a specialized large language model (LLM) tailored specifically for agri-biotechnology and agricultural sciences. Developed with high-performance graphics processing unit (GPU) computing support from the Ministry of Science and ICT, this dedicated AI model marks a major milestone in integrating cutting-edge artificial intelligence with specialized scientific research. To build an AI capable of understanding the nuanced complexities of farming, plant genetics, and agricultural science, researchers trained the model on a massive repository of 36,000 academic materials. This training dataset includes approximately 20,000 research papers collected through partnerships with five major agricultural life science academic societies, 13,000 papers from international peer-reviewed journals published by the RDA, and around 3,000 comprehensive research reports. Crucially, unlike generic consumer AI models that can generate unverified answers or hallucinate facts, this specialized model is engineered to cite clear evidence and scholarly sources for every response it provides. The new model has been integrated into the National Agricultural Biotechnology Information Center (NABIC) platform. Following internal pilot testing that began in October, the system is slated for public release starting in January, opening up vital agricultural bioinformatics and knowledge tools to external researchers, industry specialists, and university labs across the country. Looking ahead, the RDA has embarked on an even more ambitious initiative: developing a fully autonomous "collaborative AI researcher" by 2028. While the current model primarily acts as an intelligent assistant that retrieves verified answers to scientific queries, the upcoming collaborative system is designed to participate actively across the entire research lifecycle. It will analyze complex agricultural big data, help formulate testable scientific hypotheses, and assist scientists in interpreting experimental outcomes. This initiative highlights South Korea's aggressive national drive to apply advanced AI infrastructure to foundational industries, positioning smart farming and biotechnology as strategic pillars for the future.
Sponsored · Ad