KISDI Presents Data-Driven Policy Support Outcomes for Population and AI Challenges, from Forecasting Models to an Enhanced Data Platform (detail)

※ URL(Korean): https://www.kisdi.re.kr/bbs/view.do?bbsSn=115013&key=m2101113055776&pageIndex=1&sc=&sw=

KISDI Policy Report (25-12)
2025 Data-Driven Future Forecasting and Policy Support Project

KISDI Presents Data-Driven Policy Support Outcomes for Population and AI Challenges, from Forecasting Models to an Enhanced Data Platform
.................................................................................................................................................
▲ Forecasts labor supply-demand imbalances between the ICT services and ICT manufacturing sectors under demographic change
▲ Measures AI occupational exposure using large language models (LLMs) and identifies differences across occupations
▲ Upgrades KINDA into an AI-powered policy insight platform
▲ Operates the AI and Big Data Research Group to strengthen collaboration among researchers and policymakers

The Korea Information Society Development Institute (KISDI, President Sangkyu Rhee) announced the outcomes of its data-driven research and forecasting initiatives addressing population and AI-related issues, together with the advancement of its **KISDI INsight DAta platform (KINDA)** and collaborative research network to enhance the policy application of research findings.

As generative AI continues to expand, high-quality large-scale data have become increasingly important for evidence-based policymaking. This trend has heightened the need for public-sector institutions to utilize data analytics and forecasting to support policy design and decision-making.

Led by KISDI, the **2025 Data-Driven Future Forecasting and Policy Support Project** is a multi-year collaborative initiative involving research institutes affiliated with the National Research Council for Economics, Humanities and Social Sciences (NRC). The project identifies emerging policy issues and supports evidence-based policymaking through data analysis and forecasting. In 2025, the project focused on developing forecasting models related to demographic change and AI, enhancing and operating the KINDA platform, and strengthening researcher collaboration through the **AI and Big Data Research Group**.

As Korea experiences both population decline and rapid demographic change, concerns have grown over their impact on the labor market. Through the study *Employment Analysis and Forecasting Model for the ICT Industry Considering Demographic Change*, KISDI developed medium- and long-term employment projections for the ICT sector. The findings suggest that labor supply-demand imbalances are likely to be more severe in ICT services than in ICT manufacturing. The report indicates that policies promoting greater participation by young workers while making better use of older workers could help mitigate future labor shortages.

The project also conducted a study titled *Measuring AI Occupational Exposure Using Large Language Models (LLMs)*. Using occupational information from the U.S. O*NET database, the researchers evaluated AI exposure across 923 occupations through a meta-evaluation approach combining multiple large language models. The average AI exposure score was 0.402, with most occupations falling between 0.2 and 0.6. Occupations involving routine and repetitive work, such as office and administrative support, recorded relatively high AI exposure, whereas occupations centered on physical work, including construction and skilled trades, showed substantially lower exposure. The findings are expected to provide a useful foundation for forecasting future changes in job tasks and labor demand resulting from advances in AI technologies.

The project also supported collaborative research across institutions, including *Enhancing a Data-Driven Forecasting Model for Legislative Demand Related to Constitutional Fundamental Rights* conducted by the Korea Legislation Research Institute and *Developing an AI-Based Population Migration Forecasting Model* conducted by the Korea Research Institute for Human Settlements.

Meanwhile, the renewed **KINDA** platform has been comprehensively redesigned from the previous NRC Data Information System (NDIS), featuring an improved user interface and user experience. It now provides an LLM-based AI news briefing service, transforming KINDA into an AI-powered policy insight platform. In addition, the AI and Big Data Research Group, composed of experts from national research institutes across diverse disciplines, has strengthened the collaborative research ecosystem by sharing the latest developments in AI and data analytics.

Youngsun Seo, Associate Research Fellow at KISDI, noted that data- and AI-based forecasting provides an essential foundation for scientific and proactive decision-making in an increasingly uncertain policy environment. She emphasized the need to institutionalize AI- and data-driven forecasting within the policymaking process while continuously enhancing the supporting data platforms and collaborative research ecosystem.

In addition to sharing the project's research outcomes, the report recommends institutionalizing AI- and data-driven forecasting throughout the policy development process, further advancing and expanding the shared use of policy support platforms, and strengthening human resource development and permanent collaborative networks for AI- and big data-based policy research.

The report is available for download from the KISDI website (http://www.kisdi.re.kr)).

KISDI News

KISDI Presents Data-Driven Policy Support Outcomes for Population and AI Challenges, from Forecasting Models to an Enhanced Data Platform