"AI is like cholesterol, everyone hates it and at the same time it is fundamental for life"

Barcelona"Until recently, Catalonia had 27 methods for collecting medical data". With this figure, the president of the Biological Sciences Section of the Institute for Catalan Studies (IEC), Jaume Reventós, has attempted to show one of the risks we face when using artificial intelligence (AI) in current and future medicine. He did so in a debate organized by the IEC itself and ARA, in which the president of the Science and Technology Section of the IEC, Alícia Casals, also participated, and which was moderated by the head of the Society section of ARA, Gemma Garrido Granger.

As has already been repeated in the field of education, Reventós has addressed the fact that AI can be positive or negative in the medical sector depending on how it is used. "It is like cholesterol, which everyone hates so much and yet it is fundamental for life. With AI, what we need is patience", he explained. Like any enzymatic reaction, he continued, the engine is high-quality raw material, which in this case is medical data. "But currently, data is still collected and treated very poorly and, therefore, is less useful than it should be. We do not have precise data and the models do not arrive on time", he insisted. According to Reventós, to have this good raw material, it is fundamental to "all move in unison" regarding how information is collected: "The gentleman from Madrid and the one from California must name the pathology the same way we do. We must all use the same way to classify a medical pathology", asserted the president of the Biological Sciences Section of the IEC, who finds it essential that there be "a single repository" that is nourished by "scientific generosity".

Currently, one of the main risks of AI in such a sensitive area as medicine is monetization and commercialization by an industry that, evidently, seeks profits. "Medical records can be sold for 1,000 euros", he exemplified. But the speaker was optimistic about the potential of this tool: "We now have more than 6,000 algorithms dedicated solely to treating breast cancer", he exemplified. Alícia Casals also delved into this potential: "We have artificial intelligence to solve many intermediate steps until reaching teleoperation, in which it is the robot that directly intervenes on the patient", she explained.

Casals has insisted on the opportunity that these types of technologies provide to practice techniques and treatments without the need for a human patient who could be affected by these practices. "It is very useful for training. Recreations can be done and training can be carried out with digital twins –recreations of each individual– without being invasive and with augmented reality adding patient information. You can also combine physical instruments with a training program that analyzes what you are doing, evaluates you, and collects the information," describes Casals.

And, obviously, the highlight is when this technology is useful for directly improving intervention on the patient. A milestone that the president of the Science and Technology Section of the IEC exemplifies with an operation on a part of the brain in which, with augmented reality, doctors can see the inside of the skull in a projection, without the need to make an extra incision to observe. Here, Reventós even adds some hopeful examples as well. "In everything related to imaging, AI has caused a night-and-day change... Machines can detect shapes that the doctor is unable to see in breast cancer and there is also the option of not needing to insert a tube to be able to look: you throw in a capsule with a camera and there is no longer a need for the doctor to watch a 45-minute video, but rather an algorithm can already help detect where to look," she explains.

The role of the doctor

Despite this idea of a hopeful future –and a dangerous one if a good and unique method for collecting data is not found–, one of the concerns of those present in the room is to what extent AI will change their daily lives when they visit the doctor. On this point, Casals and Reventós agreed that many medical disciplines will change –"there are already surgeries that are performed entirely by a robot remotely," recalled Casals– and that some doctors will dedicate themselves to training robots. "However, the position that will never disappear will be that of the family or primary care physician, that of the professional who has direct contact with the patient," asserts Reventós: "The first job of a doctor is to calm the patient and artificial intelligence will not be able to do that," he concludes.