The National Data Agency plans to establish a 'National Data for All' system that securely connects and utilizes tax, finance, healthcare, and education data without centralizing it. The agency aims to implement a 'Hub and Spoke' model, linking dedicated networks of government data centers with a focus on its comprehensive registries of population, households, housing, and businesses.
During a report on the agency's operations for 2026, held under President Lee Jae-myung, the agency presented its plan for a 'citizen-centered artificial intelligence (AI) and data system.' The agency outlined key initiatives for the second half of the year, including the establishment and utilization of an AI data management system, the creation of data governance and hubs, and the expansion of tailored statistical and data services.
Currently, data is stored separately in various agency data centers due to information security concerns, leading to limitations in inter-agency data integration and utilization. To address this, the agency will adopt a Hub and Spoke structure, where the hub consists of the comprehensive registries held by the National Data Agency, including data on the elderly, deceased individuals, and homeowners. The spokes will be managed by various government data centers, such as the National Tax Statistics Center and the Health Insurance Big Data Center.
Notably, the system will not concentrate data in one location. Instead, it will connect major data centers through a secure dedicated network, allowing users to analyze and extract necessary information from any data center.
The agency expects this approach to achieve both 'economies of scale' and 'economies of scope' in data utilization. For instance, by linking government-wide data on debt, income, and assets with population and household registries, it will be possible to assess total household debt and identify micro-level financial risks for policy support.
To enhance security, the agency will implement homomorphic encryption technology, which allows data to be stored, linked, and processed while still encrypted. The National Data Agency plans to conduct technical demonstrations by the end of this year and establish a pilot system for homomorphic encryption-based integration and analysis next year.
The agency will also promote the opening of synthetic data for AI training, which maintains statistical characteristics similar to original data while preventing the identification of sensitive information. This aims to reduce concerns about sensitive data exposure while increasing the utility of AI training and data usage.
To strengthen the foundation for AI utilization, the agency will organize internal policy documents and international organization reports into formats usable by AI, and develop a specialized AI model to be disseminated across the government.
Particularly, the agency will gradually build AI-friendly statistical metadata to ensure that AI can provide accurate responses based on official statistics. A pilot service utilizing generative AI and data linkage technology will be introduced on the National Statistical Portal (KOSIS), enabling users to find and analyze desired statistical tables using everyday language.
To expand data services that citizens can feel, the agency will develop an AI-based system for continuous monitoring of price indicators, allowing for rapid responses to fluctuations in essential goods prices. It will also implement an automatic identification and verification system for statistical citation articles and provide information on population density predictions for holidays and events.
The agency plans to offer tailored indicators based on life stages to analyze characteristics and changes among children, youth, middle-aged, and elderly populations, supporting policy development in areas such as housing, employment, consumption, retirement preparation, health, and safety.
Additionally, the report included plans to enhance regional data to support local-led growth. This involves developing indicators for investment trends, living population, and regional industry, population, and household structure statistics, as well as piloting a regional registry linked to local government administrative data and national registries. The agency will also promote regional public-private combined data services by integrating private data such as telecommunications and card data with public data.
To improve the convenience of using the statistical data center, the agency will introduce an AI code stability checker, expanding the data analysis environment to operate 24/7 instead of being limited to business hours. Plans are also underway to expedite the provision of housing ownership, employment activity, and business registration data through provisional systems and monthly or quarterly updates.
Han Hyung-jun, head of the National Data Agency, stated, 'We will complete a national data system that practically supports the lives of citizens through AI and data innovation that they can feel.'
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.