The Korean government is launching a "data space" that enables hospitals to securely share and utilize medical data. Under the plan, original data remains stored at each hospital while only analysis results are extracted, creating an environment for multi-institutional joint research and medical AI development without concerns over personal information leaks.
The Ministry of Science and ICT (MSIT) and the National Information Society Agency (NIA) announced Sunday that they are opening a call for proposals for a medical data space pilot project. The government will select one consortium this year and provide 5.6 billion won in funding, with up to 16.8 billion won ($12.4 million) to be invested over a maximum of three years — 5.6 billion won each in 2027 and 2028 — subject to annual performance reviews.
A data space is a federated data utilization framework that allows participating institutions to conduct joint research and analysis without sending original data outside their systems. Each hospital's data is stored in separate secure cloud environments based on CSAP (Cloud Security Assurance Program) certification. Only pseudonymized data is temporarily used for research purposes and then deleted. The only outputs that leave the system are AI models and analysis results.
Similar initiatives are already underway overseas. A prominent example is the global medical data sharing federation platform operated by the Mayo Clinic in the United States. The European Union is also building and operating sector-specific data spaces based on its "Data Strategy 2020," including healthcare (EHDS), automotive (CATENA-X), and agriculture (CEADS).
To participate in the project, applicants must form a consortium comprising a general coordinating body for data space operations, at least one platform development company specializing in cloud and AI, at least five medical institutions (including at least two advanced general hospitals), and at least three demand-side organizations such as companies with medical AI models. The lead organization must be either a non-profit institution, a platform company, or an advanced general hospital. Competition is expected among consortia formed between major domestic big tech cloud companies and large hospitals.
Separately, the government plans to open a general call for proposals starting next month to plan data spaces across other sectors beyond healthcare. "The medical data space pilot will create a safe and sustainable ecosystem for data sharing and utilization," said Kim Kyung-man, Director General of the Artificial Intelligence Policy Bureau at MSIT. "We expect that expanding the sharing of high-quality data accumulated by the private sector will accelerate AX (AI transformation) and strengthen the competitiveness of the AI and data industries."