Skip to main content
Will we soon be replaced by machines?

Will we soon be replaced by machines?

The prospect that much of what humans do will be replaced by machines is nothing new. From the earliest days of the industrial revolution, there has been a massive amount of human labour that has been replaced by machine-based solutions. But many predict, we are on the cusp of another, digitally-based revolution that will see much of our professional work being replaced by artificial intelligence technologies.

In a recent Harvard Business Review Webinar, Oxford professors Richard Suskind and Daniel Suskind indicate that despite the common wisdom that doctors, lawyers and other professionals will remain relatively unscathed by AI tools, their research has led them to the opposite conclusion. They opine “that we expect that within decades the traditional professions will be dismantled, leaving most, but not all, professionals to be replaced by less-expert people, new types of experts, and high-performing systems.”1

In an interesting article by Jen Wiecszer, she discusses how revolutionary AI has already become profoundly useful. She refers to the current reality that “IBM Watson could read a patient’s electronic medical record, analyze imagery of the cancer, and even look at gene sequencing of the tumor to figure out the optimal treatment plan for a particular person.”3

The phenomenon of computer technologies rivaling a doctor’s capabilities extends to many fields. For example, in a recent article in Nature, it was reported that AI systems rivaled the accuracy of 21 board certified dermatologists in the recognition of serious dermatologic conditions.4

And it is not just doctors who are at risk of, if not being replaced, certainly seeing their work change drastically with the advent of AI. Bernard Marr, in an interesting article in Forbes Magazine, suggests that many professions are at risk, or on the cusp of a massive transformation.5 He discusses the “top ten” professions at risk and they include: 1. healthcare, 2. insurance, 3. architects, 4. journalist, 5. financial industry, 6. teachers, 7. human resources, 8. marketing and advertising, 9. lawyers and 10. law enforcement.

Einstein is quoted as saying, “Computers are incredibly fast, accurate, and stupid. Human beings are incredibly slow, inaccurate, and brilliant. Together they are powerful beyond imagination.”6

Well, artificial intelligence is bringing brand new meaning to this quote. Indeed, the professional world is rapidly and disruptively changing as the adjective “stupid” is quickly becoming inaccurate.

If you have any thoughts about artificial intelligence and the implications for the medical profession, comment on the blog, or better yet, please drop by the Macklem House; my door is always open.

Richard

  1. https://hbr.org/2016/10/robots-will-replace-doctors-lawyers-and-other-professionals
  2. http://int.search.myway.com/search/AJimage.jhtml?&searchfor=watson+as+d…
  3. http://fortune.com/2016/11/02/ibm-watson-ai-3d-printing-athenahealth-doctors/
  4. http://news.stanford.edu/press-releases/2017/01/25/artificial-inteltify-skin-cancer/
  5. https://www.forbes.com/sites/bernardmarr/2016/04/25/surprisingly-these-10-professional-jobs-are-under-threat-from-big-data/#63b1f31d7426
  6. http://www.quotes.net/quote/38086

M. Chan

Tue, 05/23/2017 - 14:25

I agree with this. Four years ago I participated in a Grand Rounds debate on “the future of psychiatry” and wondered when I would be replaced by “a smartphone”. This was after I heard a talk by a Google executive based in Ottawa.

M. Chan

Thanks Dr. Chan. I’m not sure about being replaced by a smart phone, but I think we all agree that we need to embrace the revolution that digital technology will bring us, and more importantly harness which energy and help in fashioning its use.

Richard

reznickr

Jessica Trier

Tue, 05/23/2017 - 14:27

There was a really interesting plenary at CCME 2017 by Mark Prensky about this topic, and medical education in the age of the “digital native” (a term which he coined). While we need to integrate technology into our practices, machines cannot replace the empathy that a human physician can provide

Jessica Trier

Thanks Jessica,

I guess that’s The existential debate. There will always need to be a balance of understanding the importance of the human dimension and just how much a factor it is in the delivery of quality care.

Richard

reznickr

Ross Morton

Tue, 05/23/2017 - 14:27

I have a picture of Watson in an upcoming talk on the Future of Nephrology Education which I will present at Nephrology New Brunswick in June. I also make reference to the fact that the Qualcomm Xtricorder Prize has been awarded ($2,500,000)!
We have some serious competition.

Ross Morton

Thanks Ross. I am pretty confident that we are just on the beginning part of this very steep curve of what will be understood as One of the few major revolutions of humankind.

Richard

reznickr

David Lillicrap

Mon, 06/19/2017 - 13:37

Richard – as usual, I enjoyed reading your blog this week. The relevance of AI to our futures is important to us all, and I completely agree that as physicians we should be engaged in how we can optimally integrate this technology. I have followed the recent progress of AI closely since our son (a PhD graduate from Steve Scott’s neuroscience group) joined DeepMind Google in London. As you may know, last year Deepmind published in Nature the report of their AlphaGo AI system comprehensively beating a champion Go player. With 10e170 possible board configurations it is impossible to preprogram an AI system to play this game – the system has to learn how to perform well. The incorporation of deep learning algorithms is making the concept of “general AI” a feasible reality with many diverse applications such as the dermatologic diagnostic example you provide. Indeed, DeepMind has already developed a Healthcare division and is currently working with several medical groups in London to apply this new approach to healthcare delivery. It’s very good for Canada that some of the best AI minds are working in academic groups across the country. These are very exciting times, and we should begin to ensure that our medical trainees are prepared to engage with these advances. Lastly, I completely agree that there will always be a role for physicians, but that function may be quite different from what we’ve become used to in the past many decades.

David Lillicrap

Dear David,

Thanks for your thoughtful and interesting comment. I couldn’t agree more, that we are on the cusp of a phenomenal revolution in the way we do things. I expect that much in the same way that there has been exponential growth in computing power, there will also be exponential growth in its utility and medicine.

Thanks again,

Richard

reznickr

M. Chan.

Mon, 06/19/2017 - 13:37

Last year in the UK, a researcher from Cambridge asked me, “what is the digital innovation strategy in your psychiatry dept?” And I did not have an answer. So I put this out to the Faculty. I know it is out there somewhere.

M. Chan.

M Newhouse

Mon, 06/19/2017 - 13:38

Yes, robots and computers will be important handmaidens for guiding evidence-based medicine. However, we must not forget that medical practise is an ART that hopfully makes full use of the available science!
Cheers, Michael Newhouse

M Newhouse

Trevor

Mon, 06/19/2017 - 13:38

At first glance, it’s easy to imagine that AI could come to greatly change the practice of say dermatology. And certainly, it’s likely to play a much greater role. In terms of raw image comparison, it appears A.I. can perform similarly to dermatologists.
Yet – (current) A.I. is also fundamentally limited. As described in A.I. vs M.D., in the New Yorker, it lacks any power of explanation. New pathologies and new treatments are often based on clinical insight for instance, which A.I. cannot provide.

Moreover, A.I. cannot easily ask questions along diagnostic ideas while performing a directed physical exam. If a manifestion of a particular type of pathology is sufficiently different than previous human classified occurences or if classification changes, then (current) A.I. will struggle. Finally, A.I. is not flawless. It’s accuracy in classification is impressive, but with a significant error rate that would make it very useful as an assistant.

As another example, ECG interpretation is a mature field in terms of automation, but even today appears to have limitations compared to human expertise. In fact, potential over-reliance on automation could apparently in some situations result in deleterious consequences, despite it’s potential to aid clinicians.

Trevor

I agree with all you have said. One caveat…I think we can predict that the ability for computers to learn will increase exponentially and we need to be prepared for fundamental advances in the near future.

Richard

reznickr

Nicholas Diamant

Mon, 06/19/2017 - 13:40

Richard, I certainly recognize the potential of modern technology, whether AI or other forms, that are rapidly changing the learning and practice of medicine. However as an older “old time” physician who in his semi-retirement spent clinic times here, my worries at the “human” level continued as I watched the students and residents climb into the computers. I still had to show them how to work with a patient at the human level of history and physical examination appropriate to the condition and appropriate to the real human being we were seeing. Because every patient is very different at both the disease/symptom/sign level and human level, the personal interaction in terms of medical knowledge and the real nature of the individual is still of utmost importance. The human physical, emotional, psychological, spiritual aspects, and living and working conditions, as well as life long experiences, etc, all become part of the diagnosis and management of the patient. These don’t come by asking the computer or even in totality over the internet interview. We still have to behave as doctors rather than just as technicians.

Nicholas Diamant

Add new comment

Plain text

  • No HTML tags allowed.
  • Lines and paragraphs break automatically.
  • Web page addresses and email addresses turn into links automatically.