Even with 500 innovative AI solutions,
almost none are covered by health insurance

Data fragmentation at every hospital
raises concerns over 'AI bias'

Need to prevent big tech dominance

A 'Korean Foundation Model' is urgently

Editor's Note
Korean healthcare stands at a major turning point. The shift into a super-aged society, the rise in chronic diseases, a shortage of essential medical professionals, and widening regional healthcare disparities have become challenges that the existing medical system alone can no longer resolve. Both the government and the medical community are now looking beyond the mere adoption of artificial intelligence (AI) to redesigning the entire field of medicine through medical AX (AI Transformation). The objective is not simply to supplement or replace the role of physicians with AI, but to build a healthcare environment in which everyone—regardless of region or income level—can receive high-quality care. The Asia Business Daily examines how rapidly advancing AI is transforming medical practice on the ground and how it could fundamentally reshape our healthcare system going forward.
(ChatGPT generated image)

(ChatGPT generated image)

View original image


A domestic medical artificial intelligence (AI) company poured years and massive sums into research, development, and clinical trials, eventually obtaining approval from the Ministry of Food and Drug Safety. However, the company ran up against a significant barrier when trying to enter hospital settings: there was no separate reimbursement code to claim the cost of AI-assisted care under the national health insurance. Even though the technology was proven, the so-called “reimbursement gap” served as a roadblock to the spread of medical AI, as hospitals lacked the economic incentive to actively introduce such tools.



Although the acceleration of AI-driven medical innovation has ushered in the digital transformation of healthcare (AX, or AI Transformation), many point out that the compensation system and data infrastructure remain mired in the mold of past healthcare systems. Experts broadly agree that, alongside discussions about the scope of responsibility and legal safeguards for medical AI, there must be a redesign of the current procedure-centered compensation model toward one that rewards patient-centered outcomes. They also stress the need for the country to focus its resources on developing a Korean-style foundation model to secure data sovereignty in the healthcare field.



For medical staff and industry experts working in the field, the most pressing issue for sustainable medical AX is the establishment of a predictable and stable compensation system. The current health insurance fee-for-service model was designed to reward physical acts performed by medical professionals. This structure fails to fully capture the added costs that come with AI adoption and the broader social and economic value created by AI—including improved access to care, enhanced work efficiency, and greater patient safety.



The reality is that while many medical institutions want to introduce AI software as subscription-based services (SaaS), there is no policy mechanism to factor the recurring monthly costs into patient bills or hospital claims. Over the past five years, the number of medical AI products approved by the Ministry of Food and Drug Safety has surged to more than 500—over 120 newly certified in just the first half of this year—but only a handful receive reimbursement through independent health insurance codes. Most currently survive through makeshift means such as temporary designations for innovative medical technology, selective reimbursement, research grant support, or hospitals themselves footing the bill.


[The Future of Medical AX]⑦ Overcoming the Reimbursement Gap and Fragmented Data... The Conditions for Sustainable AX View original image


At a policy roundtable on the “Spread of Public Medical AX” hosted by the Ministry of Health and Welfare last March, Professor Seongmoon Jeong of the AI-based Medical Innovation Center at Kyungpook National University Hospital called for a solid compensation system: “There must be a stable reimbursement mechanism for AI software subscription and maintenance costs, as well as for advanced algorithm development, in order for both hospitals and development companies to make long-term investments with confidence.”



The argument for fundamentally redesigning the reimbursement paradigm in step with medical AI adoption is gathering strength. Seo Junbeom, Professor of Radiology at Asan Medical Center in Seoul and head of the National AI Strategy Committee’s basic healthcare task force, remarked, “Under the current fee-for-service framework, it is difficult to fully realize the value of AI-driven medical innovation. Ultimately, the transition to a value-based compensation system is required.” This means payment would not merely be made because AI was used, but only when concrete improvements are confirmed, such as reductions in patient readmission rates or demonstrable advances in access to care and patient outcomes.



Hyunwoong Shin, head of the health and medical policy research office at the Korea Institute for Health and Social Affairs, also stressed, “What matters more than AI adoption itself is policy intervention. We need to approach this as a systemic transformation, not just the introduction of new technology.” He noted, “A wholesale jump from the existing fee-for-service system to value-based compensation is not realistic. We must design stepwise transitions—such as usage-based reimbursement, performance-based compensation, and organization-level incentives—to minimize disruption.”


An Urgent Need for a ‘Korean Foundation Model’ Directly Linked to the Public Right to Health


Another major issue for medical AX is data standardization. Currently, medical institutions in Korea each use different data structures and coding systems, so even if an AI tool performs excellently in one hospital, deploying it elsewhere requires months of additional data refinement. In this process, there is a risk of ‘AI drift’—a sudden drop in an AI’s accuracy when moved from the research lab to real-world clinical environments.



Consequently, there is growing recognition of the need to establish a “Korean-style medical AI foundation model” that can safely handle individuals’ sensitive medical data, optimized for domestic conditions. The generic large language models (LLMs) that dominate the global market are ill-suited to the specifics of Korean-language medical data, unique clinical settings, and the country’s world-unique single-payer national health insurance system.



Professor Seo Junbeom underscored this point: “It is not feasible to develop AI agents optimized for Korea’s healthcare environment and culture solely by relying on global models like Google Gemini. To avoid dependence on foreign AI technologies and to protect our patients’ data while advancing the Korean healthcare system, it is essential that we have a publicly purposed, domestic medical AI foundation model. This is more than an industrial competitiveness issue; it is a matter of medical sovereignty, directly tied to the Korean public’s right to health.”


[The Future of Medical AX]

① Three minutes from diagnosis to prescription... How AI is changing the ‘golden hour’

② "AI allows medical staff to spend more time at patients’ side"

③ Faster than the doctor’s eye... AI fills medical gaps in local hospitals

④ "Tireless AI, the best navigation tool for doctors"

⑤ "Seoul-standard care even in rural areas"... Why the government has rolled up its sleeves for medical AI

⑥ "A supercar with no road to drive on"... The reality of local public healthcare

⑦ Overcoming the reimbursement gap and fragmented data... The conditions for sustainable AX



This content was produced with the assistance of AI translation services.

© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.

Today’s Briefing