Sam Altman, the Chief Executive Officer (CEO) of OpenAI, recently visited South Korea and declared plans to develop medical language generative AI, marking the beginning of related research projects domestically this year. It is expected to significantly enhance medical efficiency by developing a foundation model that serves as the basis for AI models.
According to the Ministry of Health and Welfare and the Korea Health Industry Development Institute on the 6th, a project to develop medical generative artificial intelligence (AI) aimed at improving domestic medical services will commence this year.
A representative from the Ministry of Health and Welfare noted, "We are currently in the process of selecting the main agency for the project, and the results are expected soon," and added, "We will commence the project this year to develop a medical AI foundation model."
This project is being carried out for the purpose of promoting the development of generative AI-based medical service technologies according to the 'Medical Artificial Intelligence Research and Development (R&D) Roadmap' established by the Ministry of Health and Welfare last year.
Until now, medical AI has primarily been visual-based diagnostic assistance technology that analyzes diagnostic images to aid healthcare professionals. This technology gathers the necessary information to determine the presence of diseases based on medical images such as X-rays and computed tomography (CT) scans. In contrast, applications of language-based AI technologies, like ChatGPT, have been exceedingly rare in the medical field.
This project aims to develop generative AI based on languages, voices, images, videos, and biometric signals to enhance communication between patients and healthcare professionals and to assist in medical administration. Currently, attempts are underway primarily in some advanced countries to introduce AI into the medical field, but the technology level has not yet reached commercialization. Development of technology that can process various data from medical records, mandatory medical files, and biometric images simultaneously has not been achieved yet.
The government expects that this project will help improve issues related to the hospital treatment environment, including the 'three-minute consultation culture.'
By learning from various data generated in the interactions between patients and healthcare professionals, it is perceived that the project will enable proper treatment time assurance and enhanced communication. For instance, if a patient inputs their past medical history before a consultation, this can be automatically reflected in their medical records, or the consultation details can be automatically summarized and provided to the patient. Especially, by developing a foundation model that serves as the basis for AI models, it is expected that future universality and scalability can also be secured.
Medical AI has recently gained attention as Sam Altman, the Chief Executive Officer (CEO) of OpenAI, revealed plans to develop medical ChatGPT during his visit to South Korea. At the developer workshop 'Builder Lab' held at The Plaza Hotel in Jung-gu, Seoul, on the 4th, Altman said, "We will ensure that medical groups can utilize all OpenAI products," noting that this is one of the most promising fields for increasing the pace of innovation in the medical sector and doing good work that can save many lives.
It is not yet known how specifically Altman will develop medical ChatGPT. However, since ChatGPT is an interactive chatbot AI, it is likely to differ from existing visual diagnostic assistance AI in becoming an interactive chatbot.
Domestic hospitals are also starting to develop and implement chatbot-based AI services in succession. Hallym University Medical Center signed a contract with AI software company Konan Technology at the end of last year to establish a generative AI-based inpatient record creation and knowledge consultation platform. The five hospitals under Hallym University Medical Center generate 1.15 million medical records annually. Since it takes four minutes to create one medical record, it requires a total of 76,000 hours per year to produce all records. This aims to greatly reduce that time using medical language AI.
Seoul National University Bundang Hospital is establishing a medical artificial intelligence center and developing AI technologies to enhance the quality of medical services. Samsung Medical Center also plans to build an automated medical record filing system using voice recognition AI in collaboration with Selvas AI. Other major hospitals, including Severance Hospital, Asan Medical Center, and Catholic University Hospital, are also introducing AI services to improve medical convenience. Additionally, research is actively being conducted on using language AI for training residents and diagnosing certain brain diseases, including Alzheimer's.