Although the number of artificial intelligence (AI) medical devices that help diagnose diseases and

Although the number of artificial intelligence (AI) medical devices that help diagnose diseases and read videos is increasing rapidly, no reports of related abnormal cases have been received for more than three years. Amid the expansion of the use of AI technology in the medical field, it is pointed out that the reporting system should be checked so that potential risks such as software errors can be properly identified.

According to data submitted by Rep. Heo Jong-sik of the Democratic Party of Korea, a member of the National Assembly's Health and Welfare Committee, from the Ministry of Food and Drug Safety on the 12th, the number of AI-based software medical devices approved, certified, and reported increased from 62 in 2023 to 108 in 2024 and 157 last year.

It has increased by about 2.5 times in two years, and 55 cases were added in the first quarter of this year, bringing the cumulative number of cases to 382 from 2023.

However, from 2023 to June this year, there were no reports of abnormalities related to AI and digital medical devices in Korea.

This is in contrast to the 27 reports of abnormal cases in all new technology medical devices, including in vitro diagnostic medical devices, during the same period.

AI medical devices are used to assist in reading images such as computed tomography (CT) or magnetic resonance imaging (MRI) or to support disease diagnosis.

If algorithm errors or reading errors occur in such medical devices, there is a possibility that patient safety will be affected, such as delayed diagnosis.

The problem is that it is difficult to confirm whether there have been no actual abnormal cases in AI medical devices or whether the problems that have occurred have not been properly reflected in the reporting system.

Lawmaker Heo pointed out, "The fact that there has been no AI medical device for more than three years, even though reporting (abnormal cases) is a legal obligation, could signal that the reporting system to detect software errors is not working."

"We need to reorganize standards so that software-specific risks can be reported as abnormal cases and check whether medical institutions and companies actually report them," he said.