Food and drug safety ministry eases clinical trial rules for medical AI devices

The food and drug safety ministry said Tuesday it revised regulations to clarify clinical trial data requirements for digital medical devices.

Amending the regulation governing digital medical products on July 27, the ministry established concrete criteria for when performance verification data can substitute for clinical trial results. Officials analyzed approximately 1,000 existing software approvals to determine which device types could safely bypass traditional trials based on their intended use.

Under the new rules, software that calculates treatment locations, measures biomedical indicators from medical device data, or provides anatomical information without diagnosing specific diseases may now be approved using only performance verification materials. Devices that simply compute medication dosages within prescribed ranges or compare patient data against established clinical guidelines also qualify for the streamlined process.

The revision expands special treatment provisions for advanced artificial intelligence systems, extending eligibility from current second-grade devices to include ungraded products such as multimodal generative AI tools. The period allowed for real-world data collection under this provision increases from one year to a maximum of three years.

Kim Myung-ho, director general of the ministry's Medical Device Safety Bureau, said the changes would enhance regulatory predictability while activating product development based on the latest artificial intelligence technologies. He added that detailed guidelines for good manufacturing practice certification would be prepared soon.