Artificial intelligence (AI) is a “tsunami” that’s coming to healthcare. That’s what Naveen Jain, CEO of Viome, a small company focused on applying AI to healthcare, told CNBC on Thursday. I suspect he’s right. The waves have already started picking up force.
Three major milestones were reached related to the use of AI in healthcare in the last four weeks. There’s a good chance you haven’t heard about any of them. But just as undersea landslides that aren’t seen can lead to a tsunami forming, these three developments in November could result in a storm that changes healthcare as we know it. And it could create fortunes in the process.
Image source: Getty Images.
1. A robot passed a national medical exam
The first of these big milestones was announced on Nov. 6, when results from China’s National Medical Licensing Examination (NMLE) were released. For the first time ever, an AI-powered robot passed the medical exam. It didn’t just squeak by, either. The robot scored 456, nearly 100 points higher than the passing score of 360.
This robot, named Xiaoyi, was developed by Chinese technology company iFlytek and Tsinghua University. The organizations’ teams used AI deep learning algorithms to process data from medical textbooks, clinical guidelines, and medical cases.
So is Dr. Roboto about to begin making rounds? No, but the potential use of the technology is significant. China faces a serious shortage of general practice physicians in rural areas. While Xiaoyi won’t take the place of human doctors, iFlytek plans to launch the robot in March 2018 with the goals of assisting physicians improve their efficiency. Over the longer term, the company wants to use AI to train more general practitioners and improve cancer treatment.
2. AI beat human radiologists at diagnosing pneumonia
Just over one week after the monumental AI success in China, researchers at Stanford University in California announced another breakthrough. The Stanford team published a paper online on Nov. 14 describing an AI algorithm called CheXNet, which diagnoses 14 medical conditions based on chest X-ray images. ChexNet can even beat human radiologists at accurately diagnosing pneumonia.
This achievement was made possible by the release of a public data set on Sept. 26 by the National Institutes of Health Clinical Center. The data set included 112,120 frontal-view chest X-ray images, each of them labeled with at least one of 14 medical conditions.