Forum of the National Academy of Medicine of Korea and Korea Health Industry Development Institute
Need for Local Validation Beyond Data Concentration at Tertiary Hospitals

There have been calls to fill gaps in regional, essential, and public healthcare not just by focusing on artificial intelligence (AI) adoption, but by first establishing the necessary systems: human resources, data, and institutional infrastructure. The argument is that AI development should move beyond specialized general hospitals, building systems for creating and validating data at the local level, and simultaneously establishing front-line personnel, standards for compensation, and criteria for responsibility pertaining to AI utilization.


The National Academy of Medicine of Korea and the Korea Health Industry Development Institute held the 32nd Healthcare Forum on the 16th at the Lee Geonhee Hall of the Cancer Research Institute, Seoul National University Hospital in Jongno-gu, Seoul, under the theme "How to Operate AI Basic Healthcare in Local Communities." Photo by Jo Inkyung

The National Academy of Medicine of Korea and the Korea Health Industry Development Institute held the 32nd Healthcare Forum on the 16th at the Lee Geonhee Hall of the Cancer Research Institute, Seoul National University Hospital in Jongno-gu, Seoul, under the theme "How to Operate AI Basic Healthcare in Local Communities." Photo by Jo Inkyung

View original image

The National Academy of Medicine of Korea and the Korea Health Industry Development Institute held the 32nd Healthcare Forum on the 16th at the Lee Kun-hee Hall, Cancer Research Institute, Seoul National University Hospital in Jongno-gu, Seoul, under the theme "How to Operate AI Basic Healthcare in Local Communities." This event marked the first follow-up academic discussion since the government announced its "AI Basic Healthcare Strategy" on July 22.


In the keynote presentation, Seo Junbeom, head of the medical subgroup at the National AI Strategy Committee and professor at Ulsan University College of Medicine, explained that AI-powered diagnostic support and foundation models could enable a shift toward preventive healthcare, and outlined the tasks for AI transformation in emergency care, public health centers, hospitals, and data sectors.


The first session traced the healthcare journeys of local residents, examining both the opportunities and limitations of AI. Key discussion topics included out-of-hospital patient management, non-face-to-face healthcare, public primary care, and emergency medical services.


Heon Seong Kim, professor at Catholic University College of Medicine, proposed that legally approved non-face-to-face care should shift from a non-reimbursed, profit-driven basis to focus on chronic disease management. Seung Guk Ki, president of the Korean Association of Preventive Medicine, pointed out the limitations of AI trained on claims data and stated, "Collaboration is needed between doctors utilizing AI in clinical settings and those generating data in the community."


Geonhee Park, director of Pyeongchang County Health Medical Center, proposed field validation for AI-based primary care policies in underserved areas by referencing the achievements of frailty prevention and smart chronic disease management in Pyeongchang County. Hyunwook Ryu, professor at Kyungpook National University School of Medicine, identified the causes of emergency room crowd-outs as inaccurate severity triage, flawed hospital information, and a fragmented emergency medical system. He presented a platform that uses AI to analyze patient conditions and each hospital’s capacity, recommending optimal transfer destinations.


"A Unified IT System for Public Hospitals Needed Before AI"

The subsequent session addressed foundational infrastructure, workforce, multidisciplinary collaboration, and accountability frameworks required before applying AI to regional healthcare in practice.


Seung Yeon Jo, chief of surgery at Yeongwol Medical Center, identified structural gaps in regional healthcare: a shortage of facilities, doctorless areas, long travel distances, night and weekend service gaps, and fragmented information and capacity among institutions. He particularly emphasized that before AI adoption, it is imperative to build a unified IT system for public hospitals and secure standardized data, suggesting: "We need to establish cloud-based electronic medical records (EMRs) at regional medical centers, standardize data through a common specification, and facilitate information sharing between institutions to create the data ecosystem required for AI utilization."


Dong Yoon Kang, professor at Ulsan University College of Medicine, stressed the need for personnel who can be with patients, check information during the AI care process, and manage follow-up, particularly given the aging of medically underserved areas and their limited digital accessibility. He explained that to realize AI-enabled multidisciplinary collaboration in regional healthcare, we must design surrounding personnel, equipment, data, and workflow systems as an integrated whole.


The question of allocating responsibility for medical acts involving AI was also raised. Hyungwook Park, professor at Dankook University College of Medicine, pointed out the need for institutional standards on how to distribute accountability among medical staff, healthcare institutions, and AI developers in AI-assisted care. Wonchul Cha, head of the Digital Innovation Center at Samsung Medical Center, stressed that to make the 'AI highway' work, digital transformation (DX) and AI transformation (AX) must go hand-in-hand rather than merely introducing isolated AI technologies.


Data Concentration at Tertiary Hospitals Also a Problem... "Training and Validation Must Occur Locally"

The designated discussion session produced concrete recommendations concerning the personnel, infrastructure, data, and evaluation systems that will determine the success of AI in regional healthcare.


Sookran Jang, professor at Chung-Ang University College of Nursing, highlighted the need for expanding and stabilizing the workforce of visiting nurses in parallel with investment in AI. She stressed that AI may detect warning signs, but “if there is no one to verify and respond, AI will not work effectively in local communities.”


Suyong Shin, head of advanced technology research at Kakao Healthcare, noted that since individual hospitals cannot independently build AI infrastructure (such as GPUs and high-speed networks), government-level support is necessary. Namguk Kim, professor at Ulsan University College of Medicine, pointed out that because AI infrastructure environments differ between large hospitals and regional, essential, or public facilities, a one-size-fits-all approach is not appropriate.



Young Sung Lee, policy development commissioner at the National Academy of Medicine of Korea, proposed, "For AI-based basic healthcare to work in local communities, regional healthcare institutions must secure IT, personnel, and financial resources, with AI trained and validated on local data." He added, “Along with establishing standards for compensation and responsibility for personnel supporting patients, continuous validation and post-implementation evaluation—focused not just on the introduction of AI, but on actual patient outcomes—are necessary.”


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