It’s 2318 CE. You (or your distant offspring, if cryonics isn’t your thing) wake up on a calm Saturday morning to find that, overnight, you were joined by a most unwelcome pair of guests: a high fever, and a dreadfully sore throat. “Barnacles”, you languidly think as you lie still in bed, the sheets pulled over your head. “What a way to start a Saturday.”
It is here that you are confronted by two treatment options.
Option #1: hover (“drive”) on over to the nearest Walmart (yes, it’s still around) and find yourself a line for one of the innovative “Health and Happiness 2.0” scanners that were recently installed there. Once it’s your turn, sanitize your hands and pay a nominal fee to gain access to the scanner, which looks like a photo booth that has fallen into a pool of white paint. Step inside to find a large touchscreen which will ask for your preferred language and accent. Select “English” and “British.” A genteel-sounding voice, much like that of David Attenborough, will greet you and ask for your state-issued medical identification number. Provide the number. With this done, you will be asked what your health concern is today. Say, “My throat hurts, and I feel feverish.” “I am so sorry to hear that,” the scanner will reply. “May you be so kind as to face the screen and open your mouth?” Do as instructed. A red bar of light will then be projected from the screen and moved back and forth inside your mouth. “I am afraid that what we have here is a rather nasty case of strep throat. This is eminently treatable, however, with the right medicine. One moment, please.” A minute later, a robotic hand will give you a bottle of penicillin (S. pyogenes has––bless its little bacterial heart––failed to develop resistance to the antibiotic), and you will be given verbal instructions on how to take the pills. You will be told to come back if your symptoms are not gone in a week. You will be sent away from the scanner with a warm farewell: “Have a positively wonderful day.” Leave Walmart 30 minutes after you arrived.
Option #2: go to your nearest “oldschool” doctor. Sit in the waiting room for 30 minutes before being called back. Meet the doctor and have them manually (as opposed to robotically) shine a microbe-identifying laser into your mouth. Be told by said doctor, with a charming sprinkle of sympathy, that you have strep throat. Receive a prescription for penicillin and leave the office an hour and 30 minutes after you arrived.
Which option seems best to you, whether past or present?
Option #1 definitely has some appeal: it’s fast, efficient, and affordable. What’s more, it doesn’t require that your future self interact with a single person, which is good because you’re sick and not in the mood for human interaction.
But let’s consider the context in more depth. Imagine that humanity eventually reaches a point in its technological development at which it could dispense with all human healthcare providers if it so wanted to. There would be no nurse practitioners, physical therapists, doctors, physician assistants, and so forth; there would only be computers/machines, androids wearing white coats, and a whole lot of algorithms. It would be like Option #1, but for all age groups, for all possible medical concerns and questions, and for all nations.
My questions are: would this be a good (i.e., desirable) thing?, and do we need human providers in healthcare?
On the one hand, there are those who would answer Yes to the first question and No to the second. Their reasoning might sound something like this.
“Human healthcare providers could not be replaced by machines if there were some function which the former performed which cannot, in principle, be performed by the latter. Doctors––to single them out––engage in online charting, physical exams, ordering labs and tests, interpreting said labs and tests, making diagnoses, providing advice and motivation to patients, referring patients to other doctors or healthcare providers, prescribing drugs, teaching intimidated students and stressed out residents, publishing research studies, performing surgeries, and doing many other things related to patient care. Which of these functions, I ask, could not be performed by a machine? The answer is: none of them. While it might be difficult for a single machine to perform all of these functions, this does not change the fact that once our technology is suitably developed, machines should one day be able to do everything that human providers do. And since machines are increasingly capable of outperforming humans on everything from chess to chest x-ray interpretation, it would be quite good if machines were to replace human providers. A human provider, no matter how skilled, cannot outperform a well-programmed machine; human fallibility will make sure of that. And as we know a little too well, human fallibility in the medical arena can have devastating, tragic consequences. But this is not the only advantage computers have over humans here: computers do not grow fatigued, complain about their salaries or work hours, or show unwanted emotions. Thus, it would be for our benefit if computers were to replace human providers in medicine; the latter are certainly not needed.”
On the other hand, there are those who would answer Yes to the second question and No to the first. Their reasoning might sound something like this.
“One need not be a card-carrying Luddite to see that medicine cannot survive once its human element has been subtracted away. Are we to believe that the hospice patient dying of cancer will be content to receive their medical care from a robot on wheels, or that an anxious octogenarian about to have their gallbladder taken out will be happy to know that their surgeon’s “name” is similar to that of the Terminator’s? Absolutely not. This is because it is the distinctly human parts of medicine which keep patients sane and hopeful. A shake of the hand, a pat on the shoulder, a witty joke––these simple actions are the toast on which the butter and jam of medicine have always been spread. Familiarity fosters comfort, and there is nothing so familiar to us as our fellow humans and their normal, spontaneous behavior. What is more, it is only our fellow humans who can offer that most treasured and underappreciated of gifts: personal presence. While it would be possible to replace human providers with machines while simply hiring people with no medical training to walk around the hospital, offering, a few minutes at a time, a sense of personal presence to all patients, it is no doubt immeasurably more effective if the source of one’s medical care and the source of one’s comfort in times of medical need is one and the same. Indeed, this is why a patient can be unhappy even with a competent healthcare team and supportive friends and family: they want the people whose job it is to care for them to show signs of actually caring for them. Thus, it would be for our detriment if computers were to replace human providers in medicine; the latter are far from unnecessary.”
In this way our two interlocutors can carry on indefinitely, refusing to cede any conceptual territory to their opponent.
I smell a whiff of truth in both positions, so I propose a compromise. In my opinion, it will be good for machines to replace human providers some of the time, and humans are unnecessary to medicine some of the time. Humans will be both good and necessary in medicine where they can offer those things which machines will struggle to offer (personal presence, comforting gestures, humor, irony, familiarity, spontaneity, and the like), and where these things are desired.
Let’s return to your future self from earlier (the one with strep throat) to illustrate these points. Suppose you are otherwise healthy and are not particularly interested in chumming around with anyone while being diagnosed and treated. In this case, I think it wouldn’t be a bad thing if you were to bypass all human providers and go straight to the machines for your care. Your medical issue is rather straightforward, and you don’t have a deep longing for human interaction as you work through this issue.
Suppose, however, that you’re a bit of a hypochondriac. In this case, you may very well prefer to see a human provider so that you can receive adequate reassurance that your illness is relatively minor and should respond well to penicillin. And, as it goes without saying, you wouldn’t be able to see a human provider if they had all been replaced by machines.
Now suppose––to change the details of the illustration––that you’re a bit of a hypochondriac, and there is a nasty leukemia brewing deep in your marrow. Then you’re going to seriously want some human providers to help you through the long journey that follows your diagnosis, otherwise you’ll be at risk for having a stress-related mental breakdown. Indeed, a single hug from a human provider may mean more to you in this case than all of the pre-programmed medical management plans of any computer.
Hence my conclusion: machines can replace human providers in only some cases. Just how many cases is hard to say, but that doesn’t change the fact that I foresee medicine utilizing human providers as long as our species is still around. Just as social media websites neither can nor should replace face-to-face relationships, medical machinery neither can nor should replace face-to-face medical relationships.
But––it might be objected––who is to say that we will not someday create machines which are so humanoid in appearance and behavior that we wouldn’t be able to tell them apart from real humans? Wouldn’t we then be able to replace human providers with machines without any qualms?
Well, I incline toward the view that if we were to construct a maximally humanoid machine, there’s a good chance that it just would be a human. If, from a metaphysical viewpoint, humans are purely material in nature, then it would almost certainly be a human, and if humans are not purely material in nature (e.g., by virtue of having an immaterial mind), then it would arguably be structurally complicated enough to automatically instantiate an immaterial mind upon its completion, just as the human brain is presumably complicated enough to do the same once it is sufficiently developed in utero. Either way, the question is moot.
That’s my take on the objection, for what it’s worth.
Until next time, fair reader!