THE Artificial Intelligence is evolving rapidly and, as it is applied in all sciences, it is also applied to the various specialties of Medicine, providing the possibility not only of more accurate diagnosis and better prediction of the progression of a disease, but also of providing better treatment.

At the same time, however, its use raises concerns and ethical dilemmas for the future not only of some specialties of Medicine but also for the future of humanity itself.

The above was pointed out by the emeritus professor of Pulmonology of the University of Crete and visiting professor of the Computer Department of the University of Crete, Nikos Siafakas speaking to APE-MBE, on the occasion of his tonight’s lecture on “Pneumology in the era of artificial intelligence”, in the context of the 6th Panhellenic Conference on Thoracic Diseases of Greece held in Thessaloniki.

Will robots replace the doctor?

Mr. Siafaka’s lecture will begin with the question “will robots ever replace the doctor”.

“It is a question that concerns many people because artificial intelligence will soon reach human intelligence and when it reaches and surpasses it, man will not be the most intelligent being. It will be the machine. And the machine improves itself. But until it gets to that point it can make a lot of mistakes and the danger from artificial intelligence is great. Its uncontrolled use, its automatic use, can go wrong,” says Mr. Siafakas to APE-MPE, noting that the over-use of artificial intelligence, along with nuclear disaster and climate change, is considered one of the great risks that can destroy humanity. At the same time, he points out that there are ethical dilemmas in Medicine and Pulmonology, regarding who handles the data so that artificial intelligence can give good results.

“We have to give ‘food’ to the artificial intelligence and its food is the data we have collected from our patients. This is personal data, it’s about people. The question is how he will handle them. Ethics and data security are still big questions that concern global society and sound scary,” adds Mr. Siafakas.

Which medical specialties are “threatened” by artificial intelligence?

Mr. Siafakas estimates that with the use of artificial intelligence there is a risk of disappearing Medical specialties, first of all Radiology, Cytology, Pathological Anatomy, Microbiology, Biochemical specialties and Surgery.

“The castle of Pathology, where there is more of a patient-doctor relationship, will be the last to fall, if it falls at all. In this phase there are many gaps and the risk is high. Developments in the relationship between artificial intelligence and Medicine and especially Pulmonology are rapid, they are constantly changing. How they will affect Pulmonology no one can predict 100%. What I recommend is that young pulmonologists follow the developments, use the positives of artificial intelligence as best they can but keep their eyes open because a mistake can cost a large number. That is, I may make a mistake and kill a person, but if the artificial intelligence makes a mistake and gives a diagnosis that will concern millions of people, imagine what can happen”, notes Mr. Siafakas.

At present, artificial intelligence “right hand” of the pulmonologist

“Artificial intelligence is developing rapidly and “touches” all sciences and not only Medicine and of course also Pulmonology. Its use in Pulmonology helps in diagnosis, monitoring the patient, assessing the severity of the disease, predicting how our patient will fare. It also helps in the choice of treatment, in the changes, predicts the flare-ups that cause diseases such as asthma or Chronic Obstructive Pulmonary Disease and thus is “our right hand” at the moment” notes Mr. Siafakas. At the same time, he points out that the great achievements with the use of artificial intelligence have been made mainly in imaging methods, in lung radiology and also in functional breathing tests.

“The various algorithms of artificial intelligence have helped us a lot in difficult problems. I will give you two examples: the pulmonologist sees some small nodules in the lung and does not know how they will develop. Artificial intelligence can quickly tell us whether a nodule is benign or will become malignant. In fact, the accuracy of the results it gives us can reach 85-90%. It is a very strong prediction for us to take action. So unnecessary operations are avoided, the output of the patients is much better and the treatments are much more correct. Also from all this we draw the conclusion that the health economy is helped and such studies have been done. Also, what has surprised us is that in functional breathing tests, spirometry, the diagnosis made by artificial intelligence is better than the diagnosis of pulmonologists, that is, it analyzes the results of spirometry much better and gives us the diagnosis accurately. Something similar also happens in other diseases. In lung cancer, it helps a lot in its rapid detection through radiological diagnosis, but it has reached the point of being able to tell us the type of cancer without a biopsy. Of course, a biopsy is then performed to confirm the diagnosis. So it becomes easier to choose the treatment regimen. All these are what help us so far” adds Mr. Siafakas.