AI Medicine Amazed Me

Over the past week, a friend invited me to take part in closed beta testing of AI medicine. He invited me so eagerly because I know something about both medicine and IT, and he very much wanted me involved in the AI healthcare enterprise. My interaction with AI doctors brought about a huge change in my understanding of medicine. The part I tested was the modern medicine module — what we usually call Western medicine — covering the departments of respiratory medicine, dermatology, hematology, neurology and oncology.

After the beta, I also studied the AI-medicine-related applications in already-publicly released AI models (such as ByteDance's Doubao). The "Zhang Zhongjing" and "Hua Tuo" personas in the Doubao app (users can download "Doubao" directly from smartphone app stores) also offer TCM services, so I got to see current TCM AI as well.

Once AI doctors are online, they will greatly reduce clinical misdiagnosis rates, raise clinical diagnosis and treatment standards, and let ordinary people enjoy better — and possibly cheaper (initially even completely free) — medical services. They will also finally achieve the long-sought fair distribution of medical resources. The best medical resources will no longer be constrained by time and space: AI doctors can serve patients 24 hours a day, with no waiting and no appointment limits. Compared with human doctors, AI doctors have no emotional issues, so the patient experience may be even better.

AI doctors' clinical performance is fully the equal of human doctors. It is no exaggeration to say that the diagnostic and treatment level of current AI doctors already reaches that of a medical PhD in each specialty — and even higher. It is just that some are not yet open to the public, and the ones that are have not yet spread widely. Once AI medicine becomes widespread, its speed of evolution will be far beyond what human doctors can match.

TCM AI is the pioneer of AI medicine — perhaps because the government regulates TCM relatively lightly. The "Zhang Zhongjing" in Doubao already has nearly 1.5 million consultation records, despite not having been online for long. Even the most famous TCM doctor could never accumulate that rich a clinical experience in a lifetime, so its big-data advantage is obvious.

I consulted the TCM AI with different symptoms of over forty conditions, then reviewed its prescriptions. The TCM AI's prescriptions were reasonable and reached the level of the most experienced senior TCM physicians. Of course, on difficult questions, current AI doctors adopt a strategy of automatically advising patients to see a professional doctor. As AI doctors evolve, they will surely be able to play a greater role in this area before long.

I have studied TCM and oncology for many years. Interacting with the AI doctors, I tested them with all kinds of clinical questions; they answered rapidly, and the answers were convincing. Objectively speaking, in the fields I know best — TCM and oncology — the AI doctors' replies were no worse than my own, and some of their responses were things I would not have thought of. Other senior doctors who tested alongside me felt the same.

I deeply feel that AI doctors will, in the near future, exert a huge influence on the traditional medical model. Personally, if another epidemic like SARS or COVID were to cause medical crowding in society, AI medicine would shine. Such an opportunity would make society as a whole recognize AI medicine's convenience, efficiency and economy. At that point, traditional medicine will face an unprecedented shock, and human healthcare will change dramatically.

At present, AI medicine has two main problems: first, low penetration — only a few people understand it in depth; second, the public still finds it hard to trust AI doctors. But the next major public-health crisis will solve both problems at once. At that time, a large batch of the common diseases that currently dominate medical practice will be diverted to AI doctors. AI medicine's cost advantages are so obvious that it can significantly reduce the public's medical burden while lowering misdiagnosis; I very much look forward to this prospect.

AI medical applications may mushroom over the next few years. As AI medicine evolves and competitors vie with one another, a set of industry giants very well suited for the general public will emerge. This marriage of medicine and technology will benefit many people, but it will also deal an incalculable blow to humanity's existing medical system. So all who study medicine may face the same question: in future, how will we coexist with AI doctors without being left behind?

This week's experience gave me a great jolt. I realized that if I want to improve myself comprehensively, I must learn AI technology rather than waste too much of my life in outdated educational models of low value. AI can help us master diagnostic and therapeutic techniques more efficiently, raise our clinical level, save our energy, and help us make faster and greater progress in basic medical research.

Technological progress is the general trend. Using outdated methods to study and work will not only fail to keep us ahead in future competition but will also waste our precious time. I very much hope that today's young generation of medical learners will keep pace, make good use of new technology, improve their own comprehensive abilities, and better serve patients in future — which will also sharpen their own competitiveness.

The ancients said, "A man without foresight must have immediate worries." Medical learners should look ten or twenty years ahead. What is certain is that human medicine will not be completely replaced by AI medicine, but AI medicine will change human medicine enormously. Whether we can still meet new social needs in future depends entirely on whether we are willing to study new technology deeply and view problems with fresh eyes.

AI medicine will not solve all of humanity's diseases. After AI medicine arrives, the world's great medical problems will remain great medical problems, still requiring creative medical researchers to pursue basic research. In future, AI medicine and human medical researchers will complement each other; those who keep pace will, together with AI medicine, make greater contributions to health.

All this is no longer a question of whether it will happen, but of having already happened and when it will enter the fast lane. Young friends, I sincerely hope you will climb bravely on the road of technological innovation. Though I am already middle-aged, I am willing to surf the tide of the times alongside you.