The government unveiled an "AI basic healthcare strategy" to introduce artificial intelligence (AI) across the medical field to address gaps in regional care and the concentration of patients at large hospitals in the Seoul area. The plan would allow local public health centers and clinics to receive AI support for care, and use AI for emergency patient transport and public hospital operations to make up for shortages of medical staff.
The Ministry of Health and Welfare, the Ministry of Science and ICT, and the Presidential Committee on AI on the 22nd announced the "AI basic healthcare strategy," which contains these measures.
Kim Sang-jin, director general for policy planning at the Presidential Committee on AI, said, "In an AI basic society, the most central field is healthcare," and added, "We prepared this basic strategy to preemptively introduce AI first in the medical field, where the public can feel the impact most."
He added, "As the strategy becomes more sophisticated, the benefits people enjoy in clinical settings will grow," and noted, "The committee will play a coordinating role to maximize policy effectiveness based on interagency cooperation and budget support."
◇ AI at neighborhood clinics too… "For borderline cases, support with university hospital-level information"
First, AI will be introduced in primary care settings.
Public health centers and branches nationwide will receive a "package-type AI solution" that supports imaging read assistance, voice-based medical record documentation, and chronic disease management. The AI solutions will be linked to the Public Health Information System (PHIS) so clinicians can use them without switching between separate programs.
In underserved areas with clinics in essential departments such as internal medicine, surgery, obstetrics and gynecology, pediatrics and adolescents, and family medicine, the government will provide vouchers for a "primary care AI package." Medical institutions can choose the AI solutions they need, and the government will support the expense based on usage.
Park Jung-hwan, director of the medical AI data policy division at the Health and Welfare Ministry, said, "Clinics in underserved areas lack not only medical staff but also administrative support staff," and explained, "The primary goal is to use AI to reduce ancillary tasks in administration and clinical workflows."
He further explained, "The term 'university hospital-level AI care' does not mean a neighborhood clinic turns into a university hospital; it means that when clinicians encounter patients for whom judgment is difficult on site, AI supports them in referencing university hospital-level medical information."
In medically underserved areas such as islands and remote regions, the government will also pursue a pilot in which visiting nurses conduct joint care with remote physicians based on AI-analyzed patient data.
◇ Fewer round-robin calls to ERs… "AI recommends; clinicians make the final call"
AI will also be introduced into the emergency care system.
To fix the current practice in which EMTs make one-by-one calls to multiple hospitals to find an appropriate facility, the government will build a "tentatively named emergency integrated AI platform." By analyzing in real time bed availability, staffing, and whether definitive treatment is possible, AI will recommend the most suitable hospital.
However, AI will not automatically assign a hospital. EMTs and emergency room clinicians will decide the final transport destination.
Addressing skepticism about whether "AI recommendations alone can solve refusal to accept patients," Director Park said, "AI will simultaneously display the status of medical resources and patient needs to all stakeholders in the region, creating a structure that guides a joint acceptance decision for the community."
Park added, "If each medical institution shares its resource status by direct entry or automatic linkage with EMR, the entire region will view the same screen and collaborate to decide the destination hospital." A related proof-of-concept project is underway in Daegu.
Regulatory changes will also be pursued to allow clinicians to access "my health record (PHR)" in emergencies for unconscious patients, based on prior consent.
◇ Connecting public hospitals nationwide into one… building an "AI expressway"
In public healthcare, the government will build a "public healthcare AI expressway" that connects 17 regional responsibility medical institutions and 55 local responsibility medical institutions on a single AI platform. The project will create a foundation for responsibility medical institutions nationwide to jointly use various AI services, including imaging read assistance, medical record documentation, referral and return documentation, and medical knowledge support.
Selected this year as a national AI project, the initiative will launch a pilot in 2026 and expand to responsibility medical institutions nationwide by 2029.
Aged IT systems at local medical centers will also be replaced with cloud-based hospital information systems (HIS) that can integrate with AI. The Ministry of Science and ICT (MSIT) will invest about 30 billion won from 2026 to 2027 to first develop and validate the system at the National Medical Center (NMC) and the Seoul Medical Center, after which the Health and Welfare Ministry will roll it out in earnest to local medical centers starting in 2028.
The government will also pursue "AI-native hospitals." The Health and Welfare Ministry will build, through joint development by national university hospitals, a system that uses AI to support all in-hospital services from arrival to discharge, and the Ministry of Science and ICT (MSIT) will designate and support AI-specialized hospitals in each region based on that.
◇ Supporting development of Korea-style medical AI… "New institution-level fees to be created"
The government will also develop a "Korea-style sovereign medical AI" by leveraging data scattered across public institutions and national university hospitals.
By combining data from the National Health Insurance Service, the Korea Disease Control and Prevention Agency, the National Cancer Center, and national university hospitals, the government plans to support development of base models that domestic medical AI corporations and researchers can use. Key use cases include early prediction of deterioration in acute patients, multimodal imaging interpretation, and support for treatment and surgical planning.
Alongside this, the government will ease regulations on research using pseudonymized data and establish a shared data review board (DRB) so regional startups without their own DRB can access data.
The government will also revamp the compensation system for AI medical services.
Instead of applying fees to individual AI products or medical procedures as now, the government will introduce a new fee system that compensates at the institution level based on how effectively a medical institution uses AI. Because AI has a greater impact on overall hospital operations and reducing clinicians' workload than on individual procedures, evaluation at the institution level is appropriate, it said.
The evaluation will focus not on AI itself but on factors such as the level of data standardization at the medical institution, AI governance, clinical outcomes, and hospital operational efficiency.
The government plans to conduct fee research in the second half of this year and pursue a pilot in the first half of next year.