Artificial intelligence: a new compelling challenge in medicine according to Prof. Mantovani
Based on Prof. Mantovani‘s article on the Italian newspaper Corriere della Sera.
Thanks to some encounters in the United States over the last year I have found myself, once again, reflecting on the impact of the digital revolution in medicine. It has already changed the way research is conducted, and it will increasingly impact on medical training and clinical practice.
How digital is changing medicine
Speaking with a former collaborator of mine who is now at Verily (Google Life Sciences) and with an Italian executive in this field, I was told that “only Google Life Sciences can analyze the amount and complexity of data that you scientists produce from increasingly sophisticated analyses. Machines do and will generate the hypotheses.”
This statement was more or less reaffirmed during a recent Keystone Symposium by an American friend and colleague who is developing a new method to study immune response in cancer and to identify those patients who have a better response to the activation strategies of the immune system. At a certain point, she said, she was no longer able to analyze the data available. It was only thanks to working with a company that is a leader in the field of IT solutions that she was able to do so.
What the future holds
The picture before us is clear. Machine learning enables computers to learn from examples to extrapolate patterns, using them to make predictions and new examples. In medicine this means processing all of the scientific literature available on a given disease or treatment and offering a critical analysis back to physicians and researchers, for example with hypotheses of diagnoses and treatment.
We are facing the compelling challenge of using artificial intelligence in medicine. This challenge concerns those who conduct research and, increasingly, those who work in clinical practice. It also concerns teachers and students, and will even more so in the future. Medical students in particular will be affected as how we teach must make the most of new technologies. Humanitas University, for example, was chosen by IBM to partner in the development of a First-of-a-Kind (FOAK) program for the application of cognitive intelligence to innovate teaching in medicine. The Watson Cognitive Tutor platform gives students simulated options based on real cases so they can practice history taking and making a diagnosis in a virtual setting. Read more about the IBM Watson project here.
The importance of the human element
Making the most of the advantages of machine learning, however, can make us lose sight of the fact that this technology cannot substitute the human element; machines can be of service because they can process an amount of data that the human brain cannot, but it is the latter that must guide us and give meaning to the data analyzed.
It is not easy to get used to the fact that, today, machines are capable of making hypotheses; it is our role as humans, however, to direct those hypotheses and understand which of the many paths opened up to follow. For my colleague at the Keystone Symposium, machine learning stated what she (and many of us) already believed concerning the relationship between the immune system and cancer; it did so, obviously, because it had analyzed an enormous quantity of data.
It is up to us, therefore, to accept the challenge of machine learning without forgetting how important the human element is, both scientifically and morally, to the relationship with our patients and to the difficult choices we may face that have ethical considerations. Machines, in my opinion, will never be able to do the same.