Picture a doctors' waiting room. Cramped, packed, furnace hot. If you weren't ill before you got there, your chances of catching everything from the common cold to the bubonic plague just shot up. Then relief: the doctor will see you now. Only, worryingly, the doctor is a robot. Gleaming metal, scanning you, taking readings, running algorithms. Diagnosis complete. Next patient please.
Is this science fiction or science fact? The truth is that this technology is coming soon, in fact to a growing extent it is here already. In the United States, a robot called Watson analyses millions of patients' records, lending a synthesised hand to cancer treatment. Meanwhile Woebot, the world's first robotic therapist, conducts more than two million virtual consultations a week. While this may be elementary for Watson and Woebot, replacing skilled physicians with robots is far from an open and shut case. Should we enter a brave new world where machines take the place of doctors? For a variety of fundamental ethical, emotional and practical reasons, the answer is a resounding no. For despite some benefits, A.I. must always mean artificial intelligence
, never artificial infirmary
At the heart of all forms of medicine is the primary principle of empathy towards patients. In what world would we want to replace a GP's bedside manner with Robbie the Robot? Empathy is mandatory for doctors to ensure that they treat patients with care and understanding, and as a result, there will always need to be human to human interaction for patients. Treating people
rather than just patients, understanding the bigger picture, building trust and confidence. No matter how advanced artificial intelligence becomes, it is impossible to teach empathy. Human doctors build relationships with patients, factoring compassion as well as data into treatment plans. Ultimately, when we are ill, we prefer our hand being held by good old-fashioned skin and bone, not the touch of cold steel or a hard graphite grip.
Medicine is also as much about ethics as scalpels and stethoscopes. There is no fixed scientific answer for treating patients, especially in life and death situations. Statistics and numbers alone cannot decide the best approach to complex scenarios – scenarios shaped by factors including geography, culture, religion and beliefs. These human factors are impossible for machines to interpret, and the skill needed to navigate ethical considerations cannot be programmed. It comes from doctors' training and experience, blended with the innate human understanding of right and wrong. Unlike machines, doctors have the ability to consider not only what can
be done, but what should
But what about the argument that robots can diagnose ailments better than humans? It is true that they have the ability to process incredible amounts of data at light speed. However, despite this processing ability, or 'Deep-Learning Algorithm' to give it its Sunday name, robots cannot diagnose patients in ways that doctors can. A machine's single-track approach precludes thinking outside the box, collaborating with other professionals or even a good old-fashioned educated guess.
A recent study reviewed the treatment of 3,000 children presenting to doctors with non-serious illnesses. Six children were admitted to hospital due to doctors' intuition alone, with two saved from certain death due to unidentified illnesses. This intuition, forged from a blend of experience and hard knocks from past mistakes, allows for a breadth of diagnosis not possible by a machine.
Diagnosis and treatment in emergency situations pose another significant challenge. Machines need data, rules and time to analyse. Not an ideal scenario in blue light emergencies. Imagine a critical situation, perhaps anaphylactic shock or a heart attack. Would you want to wait while a robot churns through every piece of available information, or trust an experienced doctor to react based on human initiative, experience and controlled adrenaline?
Of course, medicine is no different to any other walk of life, so money matters. However, it remains more effective to invest in training the doctors of the future rather than invest billions trying to replace their unique human abilities with machines. While complementary investment in robotics will be beneficial, the cost of developing, maintaining and running robotic solutions cannot be made to the detriment of doctors. A recent study by the University of Stanford found that the cost of robotic surgery was $2,700 per patient more than traditional surgery. There is a risk that the drive for A.I. and robotic solutions becomes a case of the Emperor's new surgical gown. The benefits are not clearly defined or understood, yet there is a drive for development in part 'just because' the technology exists.
Do these issues mean that A.I. is not worthwhile for medicine? No. Indeed, utilising A.I. in medicine is absolutely critical, and humans working with computers will create the best outcomes for patients. Technology will allow us to train future generations of doctors via virtual reality medical theatres. Diagnosis and treatment options will continue to improve.
In 2016, surgeons at Oxford University used a remotely controlled Robotic Retinal Dissection Device (R2D2) to remove a membrane one hundredth of a millimetre thick from the retina of a blind man, restoring his sight. This incredible example illustrates that A.I. is simply the next step in the continuing evolution of medicine, similar to the discovery of penicillin, the invention of the stethoscope or ongoing prosthetics and cancer treatments. A revolution is coming. However, it must not be the replacement of doctors, but rather a quantum leap forward in the tools available for them to treat tomorrow's world.
The debate over an A.I. or human doctor comes down to whether medicine's role is to offer intensive processing or intensive care. Overall patient care can only be provided with human intervention and leadership. People who care for us need to hear and understand what we are saying, feel what we are feeling. The language of medicine can never become binary. Computers may be able to more accurately collate data or carry out incisions, but they will never replace humans' ability to think, care and feel for patients. The role of doctors will undoubtedly change to work hand in latex glove with more complex technology. But robots replacing doctors? There is no prescription for that.