TOKYO — Japan will create comprehensive rules governing the use of artificial intelligence in medical equipment, aiming to speed the deployment of tools that can improve accuracy and efficiency in a health care system burdened by an aging population.
The ministry of health will draw up guidelines this fiscal year for evaluating the safety and efficacy of AI-based medical equipment.
No such criteria now exist, so manufacturers face uncertainty on whether their device will be approved to go to market. Setting clear standards could encourage more companies to enter the field, hastening advancements in medical technology.
The new rules will also make clear that the ultimate responsibility for diagnosing and treating disease rests with doctors. This point has been seen as particularly important for AI, which blurs the line between machine and human judgement.
While adopting these tools will raise health care investment in the short term, experts predict they could reduce costs in the long run by improving efficiency.
Researchers are making strides on tools that use machine learning to identify signs of disease and locate lesions in medical images.
Information technology group NEC has partnered with Japan’s National Cancer Center on an AI system that can detect 98% of polyps — growths that can turn into colon cancer — by analyzing 30 images per second during a colonoscopy. The developers plan to begin clinical trials in 2019.
Meanwhile, electronics maker Kyocera has worked with a university to apply AI in detecting skin cancer.
But even as AI makes detecting disease easier, the ultimate responsibility for diagnosing and treating disease will rest with doctors under the new rules. This is also meant to encourage development of AI-based diagnosis.
Japan has already taken steps to allow AI developers to use personal data, including medical records, to refine their technologies.