SHARE ARTICLE

The Editors’ Musings

Tina Tan, Toh Han Chong

Tina Tan

I have very mixed feelings about artificial intelligence (AI) technology.

Lest you think I am a technophobe or a wet blanket, hear me out. While we celebrate the technological advancements that have come in recent years and the exciting possibilities of how AI can help us, I would like those reading this to ask – what price are we paying for it?

My fellow mental health professionals can agree with me (as evidenced by Adj A/Prof Christopher Cheok's article) that there are pitfalls associated with the use of generative large language models that have been trained on a vast store of medical and therapeutic knowledge: our patients become their own diagnosticians and develop the fallacy that a machine can replace a real human.

Perhaps in some ways, it can. But in many other ways, it cannot.

Innovations have an important place in healthcare. They are crucial for the betterment of patients and, ultimately, for society as well. This includes the potential that AI has in streamlining processes, and in the assisting of diagnoses and all manner of procedures so that patients can access faster and optimal care. But there remains an urgent need for a proper regulatory framework to guide the legal, ethical and responsible usage of it.

Is it not better to prevent a fire than to fight it only when it happens?

With that, I leave you to chew over this issue.


Toh Han Chong

Ai Pia Cia Eh Yia ( 愛拼才會贏 in Mandarin Chinese) is a classic Hokkien song extolling hard work, determination and resilience as ingredients for success. Embracing AI to pia (Hokkien for fight) is powerfully necessary in medicine. I believe AI will not replace doctors and all healthcare workers but instead augment and improve our work. By 2022, over half of all American doctors reported burnout due to keyboard fatigue from clicking and box-checking electronic paperwork. AI will hopefully bring doctors and health workers to a digital promised land.

Titan AI companies are employing philosophers and ethicists to train AI to do good, be good and be trusted. But does AI have soul and empathy? Empathy is the vital quality for identifying future medical students by most medical schools. Recent studies have suggested that AI can actually be warmer, more understanding and more empathetic than human doctors across multiple simulated clinical scenarios. If doctors have been on their feet all day, have yet to eat, are fatigued and then have to clerk a patient at 2 am with confrontational, verbally abusive patient relatives, being empathetic can be hard compared to a no-emotion no-burnout AI bot. But AI empathy is mimicry, not real genuine emotion. Medical chatbots, while increasingly used, have recently been found to be no better than traditional ways of getting medical information, and can even give misleading, inconsistent and inaccurate advice including wrong diagnoses as often as 66% of the time.1

Just as fire and electricity have transformed civilisation and life and yet must be handled with care, so too will AI be a force that will now be with us to infinity and beyond.


References
  1. Bean AM, Payne RE, Parsons G et al. Reliability of LLMs as medical assistants for the general public: a randomized preregistered study. Nat Med 2026; 32:609-15.

Tina Tan is a psychiatrist in private practice and an alumnus of Duke-NUS Medical School. She treats mental health conditions in all age groups but has a special interest in caring for the elderly. With a love for the written word, she makes time for reading, writing and self-publishing on top of caring for her patients and loved ones.

Toh Han Chong is a senior consultant medical oncologist and Deputy CEO (Strategic Partnerships) at the National Cancer Centre Singapore. He was a former Editor of SMA News. In his free time, Dr Toh enjoys eating durians and ice cream, reading, writing, rowing and watching films. Thankfully, the latter four are not fattening.

Tag