Korea to Upgrade Health Portal with Ontology for Accurate AI Health Data
The Korea Disease Control and Prevention Agency announced on the 1st that it will reorganize the National Health Information Portal and restructure data and open systems so that artificial intelligence can utilize verified health information. The multi-phase modernization initiative, scheduled for completion by 2030, addresses widespread concerns regarding inaccurate automated medical responses by establishing structured data relationships across public health repositories.
Key Clinical Takeaways:
- The Korea Disease Control and Prevention Agency is restructuring the National Health Information Portal into an ontology-based architecture by 2030 across three distinct phases.
- The system overhaul utilizes Retrieval-Augmented Generation principles, connecting verified public health concepts to external AI tools rather than relying on unverified training sets.
- Portal utilization climbed 44.8 percent to 1311만명 in the previous year, with page views surpassing 1억회, driving the need for scalable and accurate data dissemination.
Phased Modernization of Public Health Infrastructure
The modernization plan is structured across three sequential milestones to overhaul how health information is stored, indexed, and retrieved. The repository currently maintains 835 health and disease information guides, 5,743 multimedia assets including images and videos, 1,401 simplified medical terminology definitions, and 721 educational documents. The second phase will involve developing the core ontology system, which maps concepts, attributes, and hierarchical relationships rather than relying on traditional keyword-matching algorithms. The final third phase will deploy the resulting database to power advanced digital services for both public users and clinical systems.
Integration Architecture for Generative Artificial Intelligence
To mitigate misinformation risks inherent in large language models, the agency designed the upgraded infrastructure to support Retrieval-Augmented Generation mechanics. Instead of feeding uncurated public portal data directly into native model training sets, the ontology database will act as a verified external knowledge source. This semantic structure maps hierarchical and contextual relationships among diseases, symptoms, preventive measures, and management protocols, allowing automated tools to retrieve verified clinical guidance during query generation. Commissioner Lim Seung-kwan noted that maintaining precise knowledge content is vital to ensuring citizens can build reliable health literacy when evaluating symptoms and medical conditions.
Expansion of Application Programming Interface Access
Alongside architectural upgrades, the agency is broadening external access to its verified repository through expanded Application Programming Interface channels. Since late 2022, API access has been restricted to specific use cases, including patient education materials, academic graduation projects, and data analysis research, accumulating 482 approved requests as of September 29, 2026. Under the new expansion strategy, private portals and external artificial intelligence service providers will gain standardized access to query validated health records. This open framework is designed to ensure that commercial and public-facing health applications draw from uniform, evidence-based data sources rather than unverified internet text.
Disclaimer: The information provided in this article is for educational and scientific communication purposes only and does not constitute medical advice. Always consult with a qualified healthcare provider regarding any medical condition, diagnosis, or treatment plan.