AI Giving a Diagnosis
If you take a look at leading telemedicine services like DoctorSpring, the main attraction is specialized attention from bona fide doctors rather than getting a rapid diagnosis. Artificial intelligence can be utilized in the backend of such platforms, like pairing doctor & patient profiles, but giving medical advice for life-threatening conditions isn't something either party would appreciate.
Having said that, there are still theoretical applications in which AI software can successfully diagnose patients. In one test conducted by Stanford, a piece of software was able to diagnose skin cancer with the same accuracy as a real dermatologist.
It was not an easy feat since the Computer Science department at Stanford trained their super computer-backed algorithm with 130,000 images of skin cancer. The diagnosis was purely from a visual standpoint, and would still need to be confirmed by blood tests and a second opinion from a human medical professional.
Google had also attempted develop something similar to detect retinopathy in diabetics. Their program used a similar method of training their AI with tons of images to recognize the obvious signs of retinopathy as it develops. It was specifically designed to give remote diagnostics for a certain Indian village that lacked the personnel to deal with this specific ailment.
How Prevalent is AI in Telemedicine?
What telemedicine companies are looking for in AI isn't necessarily about giving a faster diagnosis, but rather sifting through massive amounts of data in be background. This can include helping maintain and organize patient history documents, organizing cases involving multiple doctors, and improving the communication experience between doctors and patients.
When 56 of the top healthcare companies were interviewed, 86% had admitted investing in AI software. The survey also calculated that $54 million will be spent in AI research by just these companies alone.
For telemedicine companies looking to cut corner on live chat assistants for non-essential functions, you can bet that many of them using AI chat bots. In fact, there are already several chat bots specifically designed for telehealth with close to the same functionality as human customer service representatives.
The point of telemedicine isn't really about replacing doctors with technology, but rather making communication between professionals and patients more convenient. Some of the more tedious functions of healthcare surely can be automated, but in the meantime, AI in healthcare is more of a tool than a doctor replacement. If you don't feel like going into your local doctor for non-urgent medical issues, try giving telemedicine a try.