As the predictions grow around artificial intelligence, with claims it’s primed to cure diseases, cause widespread job losses, wage wars, write poetry and kill off entire industries, a global healthcare expert has warned against believing all the hype.
During a career spent working on some of medical science’s great advances, Philips global Chief Medical Officer of Healthcare Informatics Dr Roy Smythe has seen this type of over-sell before, but the way AI is being characterised is “outlandish”.
“People assume something amazing and truly transformational is about to occur tomorrow. It is likely to happen – it’s just going to take a longer time than that,” Smythe told Healthcare IT News Australia from his home in the US.
“I think it's a combination of both the excitement around the new technology and the claims that are made by companies, plus the way the media and the general public get excited about it.”
New technologies are overestimated in the short term and underestimated in the long term, he said, with the early enthusiasm veering towards the realm of science fiction.
“I was a gene therapy researcher when I was younger and I lived through that episode of hype. So I was doing this in the early 90s and I was working in one of the first gene therapy labs in the world and we were, like, this is 30 or 40 years away from significant clinical application. And we would cringe every time there would be some article The New York Times or somewhere talking about how genetic cures for some horrible diseases were right around the corner,” he said.
“We’re like, ‘Oh my gosh, there's all these scientific problems we still have to overcome.’ I think it's similar for AI.”
It’s a conservative position for a man responsible for spruiking Philips’s AI applications, including the newly announced collaboration with Nuance, a company that provides 70 per cent of radiology reports in the US. That partnership will see the Philips Illumeo technology and Nuance’s PowerScribe 360 combine to reduce discrepancies and improve accuracy in radiology reporting with the help of machine learning.
Despite this development, Smythe dismisses claims AI is about to put radiologists out of business as “a little silly”.
“We think that artificial intelligence over the next several years will increasingly make radiologists and those doing similar diagnostic work more efficient. It will augment their intelligence; it will allow them to make faster more accurate decisions.”
In years to come radiologists and other clinicians will no longer need to perform some tasks, but the increasing application of precision medicine, growing volumes of data, and “the human discernment part of it” will ensure they can never be replaced by software, he claimed.
“If you want to move away from radiologists for a minute and think about medical oncologists who are now working with complex cancer patients trying to decide on the basis of their genomic information and their clinical information and so forth what the best treatment should be. It's going to allow those people to have the time to focus more on the person and less on moving the knobs and dials.”
So while AI is unlikely to cure cancer in the next 10 years, according to Smythe, it will drastically improve treatments.
“If you have a cancer diagnosis, the treatment that you get recommended is going to be dramatically more likely to work than it is now. Does that mean that in ten years we’re going to cure cancer using AI? I don't think so,” he said.
“It's a little bit of a crapshoot now and things like chemotherapy are being recommended whether or not it’s going to work. And I think if we can all live 10 more years before we have that diagnosis the likelihood of our treatment having an impact on our course of disease and helping maintain a meaningful life and prevent suffering is going to be much greater.”
Smythe leaves the grand proclamations to other organisations, saying, “We don't want people to be disillusioned.” He is also aware that healthcare is undergoing a messy transition period with clunky technologies, such as electronic health records, creating some cynicism among clinicians.
“While the hype is out there and people in their wildest dreams can be excited about that, a lot of people are like: ‘Technology is making my life more difficult, the EHR’s difficult to document in, I’m spending more time at the keyboard and less time with my patients, and I'm afraid the technology’s going to take my job away’.”
Smythe’s down-to-earth approach isn’t a sign of lack of enthusiasm for the technological possibilities, however, with the Philips expert more akin to a pragmatic dreamer.
“The human body is like the universe; the more we find out about it the more exhilarating it is,” he said.
“I’m very bullish, I think it’s going to be transformational, no doubt. I don't even think we know how transformational it's going to be.”
But working that out will also, no doubt, take time.