Health Ministry to Survey Medical Use of AI Across Japan

The Yomiuri Shimbun

1:00 JST, May 30, 2023

The Health, Labor and Welfare Ministry will launch a nationwide survey on the situation of artificial intelligence in use at medical institutions within this fiscal year.

The survey is aimed at identifying good examples of AI use that may reduce burdens on doctors and also bad examples of problems related to AI use. The survey results will be used to draw up relevant guidelines for medical institutions that are considering introduction of AI.

Medical institutions have started using AI for various purposes, such as to support endoscopic image diagnosis and to analyze data such as throat examination images and patients’ answers in interviews with doctors to judge whether they have influenza. Also, such medical AI systems draw on data from a vast number of cases, giving them the potential to spot lesions that might be overlooked by human eyes.

On the other hand, the accuracy of AI analysis may vary depending on the amount and the quality of data fed to the AI systems, which can lead to the risk of fatal misdiagnosis in tests for serious diseases, such as cancer. Main AI systems in medicine have been approved by the government as medical equipment and are sold to medical facilities through their manufacturers. Yet it is not clear how widely they are used or whether any problems have emerged.

Therefore, the health ministry has decided to look into examples of AI in good use at hospitals and clinics across the country and to find out whether there are any related problems, such as misdiagnosis risks. The ministry will set up a specialists’ study group on the issue and bring the advantages and problems of medical AI into focus by fiscal 2024. Effective uses of AI found in the survey will be included in the new guidelines for medical institutions.

The government’s medical strategy drawn up in 2022 called for the government to move ahead with wider uses of AI in medicine. But the Japan Medical Association then issued a statement that the largest association of doctors in Japan saw it as a problem that doctors were to be blamed for misdiagnosis by AI even when the basis of the AI’s judgment is not clear.