Knowledge has long been of the essence in clinical medicine. However, with advances in machine learning (ML) and artificial intelligence (AI) coming with increasing speed, could the basic human component of clinical medicine – the relationship of the patient with the doctor and other clinicians providing services – be in danger of obsolescence?
That is becoming an increasingly tough question to answer. Machine learning capabilities are growing exponentially with every machine training cycle,1 and data scientists today are working feverishly to improve and enhance these emerging capabilities. For example, it is onerous to teach a machine to calculate square roots for the first time, but once the machine learns, it can scale its capability via training to calculate square roots of large numbers as well.
Already, big data and advanced analytic capabilities are leveraging the genome in ways that are vastly improving diagnostic and treatment capabilities, and AI algorithms have integrated into medical information- gathering as well as other routine clinical work.
This could all be fostering an ecosystem too dynamic to even have a comfort zone.
At this point, however, many human aspects of health care appear to be safe, at least for now. Although studies are claiming that most2 skilled jobs may be completely automated by the year 20603 and the majority of healthcare jobs might at least be partially automated in the next 10 years,4 I don’t believe or expect ML or AI to replace flesh-and-blood clinicians completely.
Why? Do clinicians possess a capability that machines do not? Yes: emotional intelligence.
This to me is a positive, as the idea of machines substituting for clinicians is definitely not in my comfort zone. And I doubt I am alone in my unease. In healthcare, the human component is still vital. Think about Albert Mehrabian’s 7-38-55 Rule of Personal Communication: 55% of human communication occurs via body language, 38% through tone of voice, and only 7% through spoken words.5 Patients present not just with biometric data points, but also with verbal and non-verbal cues found in their tone of voice, the speed and rhythm of their speech, and their body language – all important elements for a good diagnosis. Relying only on words and numbers, as machines must do, would miss a whopping 93% (38% + 55%) of input – input that could make a difference in that patient’s future health.
In addition, although the pace of growth in computers’ capacity to take in and access data began to slow in 2015, it is still growing rapidly. In 30 years, the president of Japanese tech-focused hedge fund Softbank believes computers could have the computational and data-storage power equivalent to an intelligence quotient of 10,000 – about 100 times greater than that of the average human.6
All of this intelligence is bound to change workplaces – including our own – and the ability to manage intelligent workplaces is clearly going to be a future skill need. That being said, it is unlikely that an algorithm would replace humans in the management of medical practices or underwriting. Medical professionals, however, are going to have to be far more than just clinicians. To manage practices that integrate human abilities with ML and AI, they will need leadership skills, strong emotional quotients (EQs), high levels of medical and technical knowledge, and a commitment to continuous improvement of their knowledge and skills, both in medicine and technology.
The medical professionals who can make the leap not just to managing machines and humans, but also to working hand-in-hand with increasingly intelligent machines, will be the most effective clinicians and underwriters, going forward. No matter how much machines and the data they analyze and generate can help, the care of health and underwriting of lives are still, fundamentally, human-driven endeavors, and are likely to remain so.
Note: For additional reading on this topic, please visit https://jamanetwork.com/journals/ jama/fullarticle/2718456 and https://jamanetwork.com/journals/jama/fullarticle/2718457.
- Rayome, AD. MIT’s automated machine learning works 100x faster than human data scientists. Tech Republic. 2017 Dec 19. https://www.techrepublic.com/article/mits- automated-machine-learning-works-100x-faster-than-human-data-scientists/.
- McKinsey Global Institute. A Future that Works: Automation, Employment, and Productivity. San Francisco : McKinsey and Company. 2017. https://www.mckinsey. com/mgi/overview/2017-in-review/automation-and-the-future-of-work/a-future-that- works-automation-employment-and-productivity
- Revell, T. AI will be able to beat us at everything by 2060, say experts. New Scientist. 2017 May 31. https://www.newscientist.com/article/2133188-ai-will-be-able-to-beat- us-at-everything-by-2060-say-experts/.
- McKinsey Global Institute. Jobs Lost, Job Gained: Workforce Transitions in a Time of Automation. 2017 Dec. https://www.mckinsey.com/~/media/mckinsey/featured%20 insights/future%20of%20organizations/what%20the%20future%20of%20work%20 will%20mean%20for%20jobs%20skills%20and%20wages/mgi-jobs-lost-jobs-gained- report-december-6-2017.ashx
- Belludi, N. Albert Mehrabian’s 7-38-55 Rule of Personal Communication. Right Attitudes: Ideas for Impact. 2008 Oct 4. https://www.rightattitudes.com/2008/10/04/7- 38-55-rule-personal-communication/.
- Pettit, H. Robots will have an IQ of 10,000 and be 100 times more intelligent than the average human in just 30 years, says billionaire CEO of SoftBank. Daily Mail. 2017 Oct 26. https://www.dailymail.co.uk/sciencetech/article-5019265/Robots-100-times- smarter-humans-30-years.html.