Donggyun Park
Professor Frank
WRIT 1E
5 December 2018
Research Argument
Have you watched any movies in which AI robots diagnose diseases, lead surgeries
and cure human bodies in marvelous ways that we could not even imagine few decades ago?
This seems to come true in the visible future. Artificial Intelligence is one of the hottest
topics in numerous research areas these days. As a scientific method, using AI is being
adopted in medical diagnosis. With this unprecedented change happening, people have
started asking ‘Can AI replace doctors?’ and many discussions on this topic are on-going.
Every person would agree that AI will significantly affect the entire conventional medical
systems, so it must be debated by people including experts from various fields. Typically, one
side insists that the AI will completely replace doctors and the other side says AI cannot or
should not replace doctors. Here, most people consider AI for medical purposes as a
counterpart of human doctors and this tendency seems to lead the discussions to issues related
with laws, ethics and people’s awareness rather than scientific proofs. In this argument, by
employing a different perspective on this topic, I hope this proposal will give us a better or
more meaningful direction to the solution that both sides would agree.
First, we will cover the prospect of the future healthcare system in order to have a
sense for how the current healthcare system is going to be changed. So far many researchers
have studied and anticipated the change of healthcare system utilizing AI. In the close future
it is possible to connect most of the medical datasets from all over the world through cloud-
Park 2
based technologies and use them to boost diverse learning processes of AI1. There are already
many cases that have proved this network’s precision on many papers. Also, the historical
facts and the blueprints of the future medical diagnosis system are provided with concrete
figures. We can easily imagine how it will look, but many people have a skeptical view
toward this radical application. By people who disagree with the application of AI for the
medical purpose, it is said that AI is not as accurate as doctors2. From this reference, some of
them say ‘doctors’ gut feelings’ play an important role on medical diagnosis since human
doctors have abilities to take into account the overall information. However, so-called
Doctors’ gut feelings are based on what the doctors have learned and experiences from many
cases of their patients or clinical experiments, which does not imply that they are intuitively
judged. This means if AI gets trained with proper data sets (software) for enough time and
equips medical instruments (hardware) to perceive patients’ external conditions, it will also
be able to find such feeling-based symptoms from the patients. Meanwhile, they claim that
there are no well-established regulations or laws to judge when misdiagnosis happen. Also,
people’s awareness will not be able to follow the speed of changes in lots of medical
situations. Obviously these are very sensitive issues which seem uncompromisable.
On the other hand, the benefits of using AI can be easily found from a number of
articles or experiments3. The authors collected the data from various medical methods and
1
Fei Jiang et al, “Artificial intelligence in healthcare: past, present and future | Stroke and Vascular Neurology”.
https://svn.bmj.com/content/2/4/230
2
Jeff Lagasse, “AI can’t replace doctor’s gut instincts, MIT study says, Healthcare Finance News”,
https://www.healthcarefinancenews.com/news/artificial-intelligence-cant-replace-doctors-gut-instincts-mit-study-says
3
Mihalj Bakator and Dragica Radosav, “Deep Learning and Medical Diagnosis: A Review of Literature”, Technical Faculty
Based on statistical evidences, they concluded that the even wider use of deep learning is
going to be applied. Also, it can be easily found that many Deep Learning applications are
getting considerably improved and some of them are already commercialized. There is a
Image(MRI) optimization using Deep Learning techniques4. The fact we really need to be
aware of is that this techniques help doctors find symptoms that might be invisible to human
eyes while they decrease doctors’ simple repetitive labor. The process is successfully
conducted within even shorter time. This development is improving the accuracy of
However, even though we have been facing a flood of scientifically useful advances, as
the opposite argument insists, there is no perfect answer including related laws or ethical
standards dealing with exceptional cases including misdiagnosis derived from using AI on the
medical fields. For example, if AI ends up showing a critical fault that might results in a
patient’s death, who should be accused of misdiagnosis between the engineers and the
doctors caring for the patient? Therefore, AI replacing doctors completely would not be
possible unless we compromise with certain ethical issues. As an alternative, I propose that
we need to interpret the AI as one of many supportive ‘tools’ controlled by engineers and
doctors. Here are some good examples. Chatbots, analyzing pathology and diagnosing rare
diseases using facial recognition are surely not replacing doctors by themselves
independently but help doctors make a decision or find symptoms that are barely detected by
doctors’ eyes5.
4
Dongwook Lee et al, “Deep Residual Learning for Accelerated MRI using Magnitude and Phase Networks”
5
Kumba Sennaar, “Machine Learning for Medical Diagnostic – 4 Current Applications”,
https://www.techemergence.com/machine-learning-medical-diagnostics-4-current-applications/
Park 4
So far, we have gone through the pros and cons about the application of AI for
time flows. Seemingly the untouchable points are the few ethical issues and the related laws
as medical diagnosis can survive patients or it can result in medical disasters at the same time.
To get rid of these risks, the different view is proposed. We will need to utilize the AI to
handle overwhelming bio-data more efficiently, but should not burden all the responsibility
on AI because AI itself was created by human beings. As a tool, it will make doctors free
from repetitive works and interpreting vague information by using big-data analysis.
Regarding to this perspective, an incumbent physician has claimed a similar way that I want
to share6. In this essay, the physician has many people who knock his office door other than
his patients and they are a lot of software and hardware engineers who have been conducting
researches on AI at the San Francisco Bay Area (which is known as a place full of
prescriptions, adjusting and orders of medical tests, he says, “It could let me get to know my
patients better, learn how a disease uniquely affects them, and give me time to coach them
As we have looked over the advantages of utilizing AI, it will release doctors from the
most of repetitive and inaccurate medical measurements and labors as they keep being
mentioned from the beginning. There are tons of proved evidences and commercialized
applications positively supporting this argument. However, at the same time we should be
prepared to perceive radically changing phases of the reality of the medical fields and become
aware of how to properly adapt ourselves to the new environment. It is not necessary to be
fully aware of how AI algorithms are stacked up all together, but by roughly understanding
6 Rahul Parikh, “AI can’t replace doctors. But it can make them better”, https://www.technologyreview.com/s/612277/ai-
cant-replace-doctors-but-it-can-make-them-better/
Park 5
how AI devices work with doctors, definitely we will feel more comfortable and supported
by AI both physically and mentally. In this sense, we will be on the top of AI.