Artificial Intelligence (AI) in medical education : The two-tailed dilemma

Starting in 1950s, AI has bloomed tremendously in recent years. AI has been one of the most popular tools in this digital technology era, an era whereby everyone wants a piece of AI ; be it digital marketing, arts and medical education. As a medical lecturer and academician, I find it important to embrace AI rather than distrusting it wholly. However, there are some caveats.

Two-tails

During my biostatistics class, I teach my students the concept of two-tailed hypothesis. To give example, let say a person wants to know if AI can generate an image of an older version of himself. Two-tails here means it either can generate a real image of an older version of himself or an image of a completely different person. It can either go right (the truth) or left (the fallacy). Thus, the AI model can be either great or downright misleading. This has some repercussions, especially in medical education. Can we allow AI to take over medical education?

AI in Medical education : the good stuff

Medical learners can be classified into undergraduate, postgraduate and other medical professionals aimed at acquiring knowledge in medicine. Before this, there was only a notebook and a pen or pencil. Even the famous medical illustrator and surgeon, Dr Frank H. Netter, drew by hand the human body which was a remarkable feat during his time and even now.

However, have you ever thought of learning anatomy at your home without actually looking at a cadaver? A cadaver is a dead body, by the way. This concept is known as virtual reality (VR) and, to a certain extend, augmented reality (AR). You can dissect the body in real-time without the need to smell the formaline-covered cadaver. It really enhances learning and makes learning equitable.

If a professor in anatomy demonstrates the dissection in a large group, only a handful will benefit. With VR, every student can learn and do themselves.

Ethical issues of AI in medical education

The issue comes when the students wanted to submit their written essay assignments. The irresponsible ones will abuse generative AI such as ChatGPT to finish it fast. Let me give you an example. The question will be:

Explain the importance of tuberculosis prevention.

Through my knowledge at the time of writing, my answer is:

“Tuberculosis prevention is done to break the chain of transmission. It encompasses awareness on the mode of transmission, which is respiratory droplet, hence wearing mask especially N95 is a necessity. Ensuring good ventilation and avoid over crowdedness is also important. It can be controlled by ensuring patients get their directly-observed treatment, short course (DOTS) at the nearest health clinic”

If we use the same question (or prompt as they call it) in ChatGPT, the answer will be:

“Tuberculosis (TB) is a highly contagious and potentially fatal bacterial infection that affects the lungs and can spread to other parts of the body. Prevention is crucial in controlling the spread of TB and reducing the incidence of the disease. This can be achieved through measures such as:

  1. Early diagnosis and treatment: Identifying and treating cases of TB early can prevent further spread of the disease to others.
  2. Vaccination: The BCG vaccine is available for newborns and young children and can help prevent severe forms of TB.
  3. Health education: Raising awareness about TB, its transmission, and symptoms, can help people take appropriate measures to prevent infection.
  4. Infection control: Implementing proper infection control measures, such as wearing masks and ventilation, in healthcare settings can prevent the spread of TB.

Preventing TB is crucial for reducing the burden of disease and improving public health, especially in countries with high rates of TB incidence.”

Hence, as you can see the answer is human-like and quite comprehensive. There is a move in academia to use software to check authenticity such as using Turnitin’s AI detector or GPTZero.

How to solve this two-tailed dilemma?

  • Learn and relearn : There is no better way but to learn the basics of AI and not just be an end-user. There are a multitude of online AI courses which some are free and can give certification such as in Udemy and Coursera. You will be able to differentiate between a good and poor AI model.
  • Practice what you learn : If you are into coding or programming, you should try Python’s Jupyter Notebook. It is a powerful yet simple platform to let AI do its magic. Do not worry, AI will not take over the world. After all, it is a software made by humans.
  • Seek correct AI mentors : In medical education, you need to talk to medical professionals who understand both medicine and the AI language. Medical statisticians and epidemiologists who work with data a lot will be good mentors, to start with.
  • Refer to ethical guidelines in AI : AI should only be a clinical decision support system and not to replace doctors as the best practitioner to give medical advice. After all, health is a matter of life and death.

Join us at our annual Healthcare Revolution Conference & Exhibition (HealthRev) where we have a plethora of topics for addressing all your clinical/non-clinical questions and opportunities! Can’t wait? Check out our upcoming workshops and courses at Disruptive Doctors Academy also!

Be part of the Disruptive Doctors Community by getting your premium membership to get full access to all our articles and services like our career coaching, opportunities to connect with like-minded doctors, and discounts on all our workshops and events! Be part of the movement and let’s disrupt health together!

Wanting to explore the options you have out there? Do download Dr Selina’s Switching Careers for Doctors eBook here to get you started on your journey and don’t forget to catch our podcast and videos on our Spotify and YouTube channel

Check out other articles: