Artificial Intelligence and Machine Learning: “Am I a real Boy?”

This article was exclusively written for the Sting by Mr Ahmed Rafay Afzal, a medical student from King Edward Medical University, Lahore, Pakistan, currently pursuing a career in United States. He is affiliated to the International Federation of Medical Students Associations (IFMSA). However, the opinions expressed in this piece belong strictly to the writer and do not necessarily reflect IFMSA’s view on the topic, nor The European Sting’s one. In the long forgotten lores of eastern medicine, the doctors were called “Hakeem”. I happen to be a descendant of one of those hakeems. The word “Hakeem” comes from the Arabic word “Hikmat” which means wisdom or all-knowing. Unfortunately, I a doctor of real medicine, do not enjoy that status anymore as compared to my great-great-grandfather who was a “Hakeem”. Today whenever I interact with patients more often than not I encounter questions like, “I read about this on the web…”, “I hear what you are saying but I think you’re wrong because medscape says otherwise…”, “I googled my symptoms and I think I have cancer…” The point being that the state of medicine has transformed from witchcraft to  scribbled prescriptions to Electronic Health Records. We have made giant leaps forward. We used to develop medical protocols keeping general principles and populations in mind, but with the advent of digital health, precision medicine has been making the rounds. All this happens by the magic of Artificial Intelligence, which makes the patient the point of importance, creates a large amount of statistical data about an individual and gives medical professionals tools to sort through and analyze that data. With such huge amounts of data it has almost become impossible for a physician to analyze it. This is where machine learning comes into play, although not objectively artificial intelligence, machine learning is the still the biggest arsenal available for the physicians who are embracing the new onslaught of technology in the medical field. National Institute of Health defines precision medicine as an emerging approach for disease treatment and prevention that takes into account individual variability in genes, environment and lifestyle for each person.¹ This approach involves algorithms which use supercomputers to mine data with the help of machine learning and deep learning. These algorithms can not only generate patterns which physicians can use in their practice they have also been applied in making diagnosis almost as accurately as doctors in the field of cardiac sciences.² Same goes for radiology where machine learning algorithms can accurately pick out pathologies in radiological scans. When combined with the results of a pathologist the success rate increased upto 99.5%.³ Machine learning softwares have been written that help record the patient-doctor encounter, hence reducing the workload of the physician by writing his notes for him.With the current status quo of machine learning it might not be sufficient to replace a physician but there is enough evidence to argue that machine learning can definitely supplement a physician even in its current relatively untested phase. It’s not all good news however, A.I. has its technological and medical limitations. A computer cannot do a physical exam, it doesn’t have the cognitive ability of a physician, it doesn’t have compassion and it can’t feel for the patient. There is also an ethical consideration to be made. Just like humans, A.I. is also flawed, who is to blame when the A.I. makes a wrong decision regarding patient care. Would the society ever entrust a machine to deal with the intricate details of health records including sensitive data e.g sexual history or HIV status? Would a human ever be comfortable giving a medical history to a machine? Is there a place for replicating empathy in healthcare? How would the paradigm of human-machine interactions work in healthcare setting and what laws would govern that interaction? These are just a few of the questions that we haven’t answered yet. Even though in other industries the use of A.I. like for example in autonomous cars, where Rogers K. Germany has developed the first ethical guidelines for driverless cars.⁴ There is also another issue with misconceptions and over-exaggerations about the potential of A.I. Artificial Intelligence is revolutionary, no doubt, but its not the answer to every problem that we have in healthcare. However one thing is for sure Artificial Intelligence will not replace physicians, however, physicians who employ A.I. will replace those who don’t. References: 1.(Collins F Precision Medicine Initiative | National Institutes of Health (NIH) [Internet]. National Institutes of Health. 2015. Accessed online on the 25th of April, 2018 from:https://www.nih.gov/precision-medicine-initiative-cohort-program) 2.Luo G, Sun G, Wang K, et al. A novel left ventricular volumes prediction method based on deep learning network in cardiac MRI. Comput Cardiol [Internet]. 2010;2017:2–5. Available from: http://www.cinc.org/archives/2016/pdf/028-224.pdf 3.Wang D, Khosla A, Gargeya R, et al. Deep learning for identifying metastatic breast cancer. eprint arXiv:1606.05718, Publication Date:06/2016. [Google Scholar] 4Available from:https://motherboard.vice.com/en_us/article/599wnz/germany-has-created-the-worlds-first-ethical-guidelines-for-driverless-cars [Google Scholar] About the author Ahmed Rafay Afzal is a medical student from King Edward Medical University, Lahore, Pakistan, currently pursuing a career in United States. His primary field of interest is pediatrics. His current focus is research in pediatric Gastroenterology and on revamping the healthcare system of Pakistan with introduction of digital technology.