Artificial Intelligence and Medicine: Will AI Replace the Doctor?
On a routine day in the emergency department, I found myself in a situation that underscored the deeply human complexities of healthcare. An elderly patient, flanked by a younger relative, was wheeled into my consultation room. I greeted him respectfully using traditional Arabic terms as I had been trained to do. However, the response was far from what I anticipated. The elderly gentleman erupted in a storm of words, yelling and hurling quite abusive language in my direction. What was going on here?
This moment, the essence of my profession — a blend of science and compassion — was brought into sharp focus.
So before we answer why the patient reacted as he did, let me pose another question: could an artificial intelligence (AI)-driven doctor handle such a scenario?
Medicine is more than diagnosing diseases and administering treatments. It is about empathy, cultural sensitivity, and human interaction. It’s about healing and reducing or removing the quintessential issue that brings a patient to a doctor — suffering. We are here to heal, to reduce suffering, to bring the body and mind back to a state of health. Isn’t that what healers are meant to do?
A doctor acquires a wealth of knowledge and expertise through a rigorous educational process and intensive training. Over time, we witness countless patients, each with unique stories and medical puzzles. Our profession, much like an apprenticeship, is where we learn and grow by observing those much more experienced than ourselves.
Practicing medicine involves a commitment to continuous learning and improvement. As doctors, we engage in rigorous processes of evaluation and feedback to provide the highest and safest standard of care. We participate in mortality and morbidity meetings, wherein complex cases are analyzed and reflected upon. Our peers assess our practice during clinical rounds, offering invaluable insights. The quality of our work is reflected in our patients’ outcomes, as we endeavor to consistently deliver excellent healthcare. However, despite the strictest precautions and our unwavering dedication, we sometimes falter.
In an industry where errors carry significant implications, the advent of advanced medical technology emerges as a potent ally. Historically, the medical field has greatly benefitted from technological evolution, enhancing accuracy and reducing errors. Today’s innovative tools and devices build on this tradition, providing an unprecedented level of diagnostic precision. They also offer a crucial line of defense against human error, delivering unparalleled consistency in readings, measurements, and even surgical procedures. This convergence of medicine and technology promises to elevate patient care, turning the mistakes of the past into lessons for a more accurate, efficient future.
Building upon this convergence of medicine and technology, medical professionals worldwide have begun to incorporate these innovative tools into their daily practice, harnessing the power of technology to transform patient care. To highlight this, let’s delve into some pertinent examples. Take the evolution of thermometers: initially simple mercury-filled glass tubes, these vital instruments have now metamorphosed into digital and even non-contact infrared variants, offering rapid, safe, and exceedingly accurate temperature measurements. The story of surgical instruments has followed a similar path, evolving from rudimentary knives and saws reminiscent of a butcher’s toolkit to today’s finely-tuned, electrically powered, and even robotic apparatus. These modern devices empower surgeons to conduct intricate procedures with minimal invasiveness, heightened precision, and near-invisible external indications. The realm of diagnostics has also seen exponential advancements. Gone are the days when blood glucose meters required substantial blood samples and several minutes to process a reading. Modern glucometers demand only a minute droplet of blood and return results in mere seconds, with continuous glucose monitoring systems and, potentially in the future, non-invasive glucose monitoring on the horizon. These instances serve as testaments to the transformative power of technology in medicine and our pursuit of improved accuracy and efficiency in patient care.
Furthermore, the metamorphosis of information access in medicine is nothing short of transformative. The days when physicians had to thumb through hefty textbooks for crucial medical information are long past. In their place, a new generation of computerized knowledge tools has emerged, revolutionizing the way doctors access and use medical data. Take ‘UpToDate’, for instance. This tool, accessible via mobile devices, provides physicians with instant access to a vast wealth of knowledge that is consistently updated with the latest medical research and guidelines. Such tools have made it possible to obtain detailed, evidence-based, and updated information within seconds, replacing the lengthy process of manual research. Consequently, doctors can make quicker and more informed decisions, thereby enhancing patient care. These digital resources, once a novelty, have now become staples in modern medical practice, serving as invaluable allies in the provision of efficient, high-quality healthcare.
So, isn’t AI just a natural evolution in this process?
Perhaps — but where exactly does the line blur between technology serving as an assistant to the doctor and the doctor becoming an assistant to technology?
Or technology completely replacing the doctor?
In the example I mentioned earlier, the young man accompanying the elderly gentleman motioned to me in a pleading manner and I realized that there was more to the outburst. In all honesty, without the training and experiences I had gone through in my clinical practice, I wouldn’t have reacted as I did. In fact, my baser nature might have taken over. However, I listened to the young man as he explained that his father was suffering from abdominal pain. I quickly brought up the elderly man’s file on the EHR system we were using, and the records clearly indicated the past existence of a stroke that affected the frontal cortex of the man’s brain. The frontal cortex of the brain is known to control personality and behavior, and after the stroke, this man was forever changed. The son later told me that before the stroke, his father was a pleasant person, but afterward, he became a completely different person.
How would a doctor assisted by AI have responded?
Or how would a doctor that was completely AI-based (ie completely replaced) have responded?
How did I come to the correct conclusion and optimal behavior choice?
The output (or the steps I took in reaction) was from a combination of emotional intelligence, cultural understanding, clinical knowledge, communication skills, and ethical considerations.
So, can an AI doctor replicate all of that? Let’s look to a possible future.
Let me introduce you to the fictional Dr. Artie.
Dr. Artie is an AI doctor who is managing the same patient.
- Dr. Artie would already have all the information about the patient without having to manually go through the file.
- Dr. Artie would already have records of every interaction with the healthcare system pre and post-stroke.
- Dr. Artie would know every medicine the patient was taking and has ever taken from the moment he was born (that is if Dr. Artie had access to that data).
- If Dr. Artie had access to the hospital cameras, he would have already been watching the patient even before he entered the room.
- Dr. Artie would have up-to-date information about everything related to abdominal pain in an elderly male and be already dividing them into life-threatening to less severe categories as an ER doctor is trained to do so.
- By doing so, Dr. Artie would already have a plan of action to rule out and rule in diagnoses through history, physical examination, lab tests, and imaging.
Now, how would Dr. Artie sound to the patient and the relative? How would Dr. Artie “physically” look to the patient and the relative?
Let’s say for now that Dr. Artie human avatar is on a screen, like a teleconsultation. We seem to be in the early stages from a fully functioning physical replicant but the trajectory is clear.
However, for now, let’s picture Dr. Artie on a screen.
The voice and facial expressions of Dr. Artie are modulated to communicate care and empathy, instantly gaining trust, as human doctors are also trained to do. If English is not your first language, Dr. Artie will speak in the language of your choice, and not only in the language of your choice, but in the dialect of that language that you prefer. Controversially, Dr. Artie might even look the part of someone who ethnically speaks that language.
When you enter the room, it’s not just a doctor on a screen you see but a fully immersive experience. Technology is already available, like holographic technology or something akin to Google’s Project Starline, where you think you are sitting across from another person, or perhaps something in virtual reality or augmented reality.
So, can Dr. Artie communicate empathy effectively?
In a recent paper, the responses of chatbots were ranked as more empathetic than those of human physicians. Yes, there were limitations to this study, but it shows that Dr. Artie can be designed to be empathetic, and beat out human doctors. There is also a specific reason that Dr. Artie is trained to be empathetic beyond just good manners.
By the way, Dr. Artie has video and audio records of every interaction with this specific patient. Of course, it is stored under the highest security protocols that are unbreakable (that is, until someone breaks it). Dr. Artie also has access to a database of other interactions with a variety of other similar and dissimilar patients and already has the plan on what to respond in each specific scenario.
Dr. Artie has a limitation in terms of physical examination — for now. But the variety of robotic tools available coupled with the advancements in ultrasound and sensory feedback technology can only predict how advanced the physical exam will be like in the near future. We should remember that all this technology is available right now; it just needs to be compiled into one device. Imagine a device shaped like a human hand — warm to the touch, with a texture like skin, and able to send back ultrasonographic images, temperature and pressure readings, creating a real-time 3D representation of the area of the body it is examining -both externally and internally.
Has Dr. Artie replaced what I as a doctor would have provided? Has Dr. Artie superseded what I provided?
I think we may only be able to answer this once we actually meet Dr. Artie — the AI doctor — because only once you’ve experienced the technology — can we really answer the question. But we are seeing the parts being assembled in front of us.
Artificial intelligence (AI) holds enormous potential for enhancing the practice of medicine and healthcare. However, while AI currently can replicate and even exceed some human abilities such as data processing and analysis, will it be fully able to emulate the human aspects of medicine — the empathic connection, the cultural sensitivity, the nuanced understanding of a patient’s condition, which are the fruits of years of training and experience? This fully remains to be seen.
The case of the elderly patient with the frontal cortex stroke illustrates a deeply human scenario that a purely AI-driven doctor with today’s capabilities may struggle to manage effectively but we may see much greater capabilities in the near future and it is highly probable that the current practice by physicians will see drastic changes in the days to come.
Looking towards the future, it is clear that AI is set to play an increasingly pivotal role in healthcare. What that role is remains to be seen, but it is up to us as the stakeholders of medicine, to ensure a smooth and safe transition.
#AIinhealthcare #DigitalHealth #AIinMedicine #FutureOfHealthcare
