I would like to start this discussion with an analogy statement from Baldwin (2019):
"one should not imagine AI as a 'baby farmer' who will one day become a full-grown, superhuman farmer; a better comparison would be with the tractor. this machine is superior to humans in many areas, but cannot replace them and remains their tool"
There are many areas in medical diagnostics where AI applications can be seen, such as Data analysis and pattern recognition, telemedicine, medical imaging and so on. Nevertheless, it remains weak for the time being because it cannot do comparable things in the context of complexity, ambiguity, and contradiction.
The basis of AI is machine learning (deep learning) which requires large amounts of data and extremely powerful computers. It would require a huge amount of wealth and machines to feed complex data to build a systematic AI program that can replace medical doctors and practitioners. In the coming time, it does not seem that replacement will happen.
In my opinion, in the next 10 years, AI will not be able to completely replace doctors. It is possible that AI will be entrusted with the autonomous performance of some medical functions (for example, the X-ray interpretation), but the results of its' work will still be controlled by doctors.
In my opinion, this is the most likely scenario, based on the current situation in the world. With the appearance of "black swan" (like the COVID-19 pandemic), which will lead to a sharp increase in the morbidity incidence and burden on medical staff, humanity may be faced with a choice: to receive at least some medical assistance from AI or not to receive it at all...
1. The patterns and presentations of diseases are always changing. Sometime Experts become confused to diagnosis.
2. In future cases of "Rare diseases" will rise
3. Number of cases on multiple NCDs is rising
4. Mechanical error cannot ignored
beside above-mentioned there were other issues will discover day by day.
Best is NOT ALLOW AI TO DIRECT HEALTH SECTOR
Next, can use AI to assess the stander only. For example, X-ray lung : AI will said the normal things only that we study in medical book. Doctors will assess abnormal X-rays only. Then risk for patient will reduce.