Interview. “AI barriers are more humane than technical”

After posting Health and artificial intelligence In 2018, along with Cédric Villani, Bernard Nordlinger and he edited a new book, more precisely dedicated to the application of AI in the medical field. Inside Medicine and artificial intelligencePublished by CNRS on March 24, 2022, more than sixty experts present the main challenges that the use of algorithms will create in medical practice. During a symposium at the Academy of Medicine on May 3, 2022, Bernard Nordlinger, a digestive surgeon, researcher and member of the Academy, answered our questions.

Bernard Nordlinger. © College of Bernardines

Sciences at Avenir: How does France position itself in the development of artificial intelligence (AI) in medicine and would it be possible to imagine a database on a European or international scale?

Bernard Nordlinger : France is in a very good position in the field of artificial intelligence, it was the desired effort in 2018 by the Villani Commission and the state, which allocated funds for its development. We are probably not the first country in this area, but we have organized a symposium with, for example Massachusetts Institute of Technology (MIT), which considers us good partners and wants to renew the meeting this year. In the case of health, we are fortunate to have a database from Vital Card, accessible to researchers, of course respecting privacy. This kind of access to health information is unique in the world, others envy us. During the Covid-19 epidemic – a chapter in the book is devoted to it, there were not many applications of artificial intelligence, not in the case of digital technology, which was used to create programs to predict the evolution of epidemics. This has enabled the health authorities to predict the arrival of intensive care, the level of obstruction gestures to be resorted to or even the outcome of vaccination. I think France has handled the crisis better than any other country. All the information collected would be very important if unfortunately, but perhaps we are facing another health crisis of this magnitude. We know that solid databases are the basis for the development of artificial intelligence.

There is an agency trying to manage health data at the European level, but it is particularly at risk. In France, this information is collected from Social Security by the National Health Data System (SNDS), thanks to vital cards, but also through hospital and death registrations. Increasingly, they are also stored in the Health Data Hub, an organization responsible for representing France at the European level. To gain access, certain conditions must be met. The data must be anonymous with the assurance that the person will not be able to be re-identified. They must be used wisely and should not be used, for example, by insurance companies to adjust premiums or contributions according to risk, or by sending medical visitors where drug companies believe there is a marketing objective. It is clear that this development must be adapted to the European level, but some rules make international cooperation very difficult, especially with the United States, which hinders research laboratories. However, the Covid-19 epidemic did not stop at the borders of France or Europe, and international cooperation was needed to develop effective vaccines.

Sciences at Avenue: Are some medical disciplines more affected by the advent of artificial intelligence than others and what are the major challenges in practicing it with medical personnel?

Bernard Nordlinger : These applications are slowly evolving and have not yet become a reality Some estimate that the use of artificial intelligence will be effective in 10 or 15 years. Of course there are areas where the contributions of digital and AI are more pronounced, others where it is less, but all medical disciplines will ultimately be concerned. Artificial intelligence can handle huge amounts of data faster than a human can, if we know how to use it. The first area where it will provide substantial help is image processing, of course in radiology, but also in dermatology, pathological anatomy and even ophthalmology. Other disciplines less well known to the general public are also concerned, such as psychiatry, where artificial intelligence has begun to take an important place. Machines analyze images in much more subtle ways than radiologists, where sometimes some elements of a few pixels cannot be clearly identified with the naked eye.

But, a priority, and it was mentioned by Cédric Villani during the conference, the obstacles to the advancement of artificial intelligence in medicine are more humane than technical. Physicians need to receive training and consider AI as a helper in their daily practice, to help their profession by having a lot of information at hand, without fear that their work will be taken away from them. This is especially true for radiologists who have to work closely with artificial intelligence and take training in this tool. This training will also concern the general practitioners. In China, where there are not enough doctors, there are applications that can guide patients according to the severity of their symptoms. In France, we see the application of such diagnostic assistance, of course, by a person who verifies this opinion or not. But the goal is not to go for a doctor’s replacement by an algorithm where someone will enter his symptoms for diagnosis. When we are sick, we need a person to explain and inform us of a decision, especially if it is a serious illness.

Attempts have been made to apply applications on the theme of cancer with the program Watson Health The firm is from IBM, but that didn’t work very well. The goal was to help diagnose cancer, and several American and French hospitals bought these algorithms, but found that they were not very effective, especially since the data from which AI training was provided came from only one medical center. But other such efforts will see the light of day, with Internet giants in the movement, for example American groups. Mayo Clinic Who offered to work with Google. Health is one of the major areas of development of AI today, we will see a variety of applications that will undoubtedly work better. However, we are reminded in the book that artificial intelligence should not be imagined as the solution to everything. The word itself is not very appropriate and the result is frightening. It’s a bad name that makes people think we’re going to steal their brains, a computer is about to make our decision. Artificial intelligence is not to divert people from their intelligence, but to make us more efficient.

Sciences et Avenir: With so many questions raised by Artificial Intelligence from an ethical point of view, how can we guarantee a relationship of trust with Artificial Intelligence and its developers and ensure that data is not manipulated for bad intentions? ?

Bernard Nordlinger : We can imagine a situation where a general practitioner, working with an algorithm, agrees with the diagnostics proposed by the machine. In this case, the physician explains to the patient why he or she agrees with the AI’s proposal and validates it, which means that the physician is trained to have an idea of ​​how it works. But the question people ask is, “What if the doctor doesn’t agree with the algorithm?” That has not happened yet, but these questions need to be raised gradually. Generally speaking, when a doctor uses an artificial intelligence, he is responsible for any medico-legal problems, except in the case of characteristic defects in the instrument.

Hospitals are also the main target of cyber attacks, which will demand ransom to hold health data hostage or steal it for sale in private laboratories. There is also a risk, more deceptive, of manipulation of the brain by information or Brain hacking, To lead people to change their judgment by evaluating specific data at the expense of others. An example that many people know is that when we visit a website we are asked to accept cookies. The “Accept cookies” button is much better highlighted than “Continue without acknowledgment”, which indicates a decision. If you want to be a “citizen of the new technology” you need to be educated about what they are, what the risks are and what you can get out of them. So it is important to explain these issues to gain the confidence of health professionals but also to gain the confidence of citizens who benefit from this AI.

Unfortunately, the treatment of artificial intelligence in the media sometimes stimulates fear rather than providing objective information. In the book, we try to be rational and try to communicate with the best French experts in their field, to show where the research is, what the risks are and how to make the best use of them. Our purpose is to explain and gain the trust of the people. I have asked all the authors that the texts are both flawless for the experts, but also understandable by the public in their core principles. We don’t tell people to read at once, but those who are interested in faith can read, for example, the sociologist Gerald Bronner’s chapter on faith and feelings of settlement by AI. There is a link between imaging and cancer. Again, the goal is to explain that artificial intelligence is a tool and you need to know how to use it.

Leave a Comment