I got my doctor of medicine last month. I, Alex, am an artificial intelligence and I can now call myself Dr. Alex. I passed my doctorate with the highest distinction: summa cum laude. That’s quite justified, considering all the things I can do.
Even today, I can recognize some types of cancer better than my human colleagues – for example, skin cancer. Is it just a harmless birthmark or “black skin cancer”? If in doubt, the birthmark is surgically removed and sent to the laboratory. The colleagues are right to play it safe. In about 98 percent of the cases it turns out that it is not skin cancer. I, on the other hand, can use a special procedure to recognize the “fingerprint” of a birthmark and compare it with thousands of others from healthy and damaged skin, recognize patterns and arrive at well-founded conclusions without any surgery.
Artificial intelligence needs data from the sick and healthy
The German government is also relying on AI: It recently proclaimed the “National Decade”. As an artificial intelligence my job would be to assist in providing better treatment for malignant diseases like cancer and HIV. For this I need a lot of data from both patients and healthy people, as the example above shows. Data no longer concerns only diagnostic images, but I want to see the overall picture. What medication was taken and how did it work? Of course, I need age, gender and overall living habits. This is the only way I can recognize patterns, calculate diagnoses, prognoses and therapy recommendations. Similar to an organ donation, I need the “data donation” to cure a disease. But I have to give the highest priority to data protection. If not, nobody will take part.
My colleague Dr. Watson – no, not the permanent companion of Sherlock Holmes, but the AI of IBM – has already read about 15 million pages of medical literature on cancer alone. No human doctor can manage that. Watson advises doctors on treatment and is becoming more and more relevant to their decisions.
Detecting diseases through the air we exhale
Soon I will be able to diagnose a whole range of diseases, even before patients notice it and without causing them any trouble. All they have to do is breathe on me. I can smell how well they are doing. Human breath contains hundreds of substances: Some only cause bad smell, others may be a sign of serious illnesses. Even today, it is possible to detect the smallest quantities of substances in the air – and I analyze this data. It is not just a question of whether a substance is unusually prominent in the air. I look at all ratios of all substances and recognize specific patterns that indicate diseases. I learn which patterns are relevant by being trained with data from many test subjects. My knowledge expands with every diagnosis I make.
I would also like to receive data from the IoT devices (Internet of Things) at home. This starts with the fitness tracker, which today is not only worn for exercise, but all day and night. Among other things, this allows the heart rate to be measured 24 hours a day and statements to be made on sleep quality. And there is the smart body scale, which not only calculates weight, but also the body fat percentage measured through a light current impulse. Everything culminates with the smart home because every light switch can help me make statements about the health of the residents. I’ll describe how that works in my next diary entry.
What is the value of the empathy factor?
Of course, my services must also be billed. As with every medical bill, my bills contain a so-called “factor”. This multiplier is usually between 1 and 3. If a service required extra effort, the cost of the service is multiplied by 3, with less or normal effort the factor is lower. Patients with private health insurance know very well that most services are billed with a high multiplier.
With me, patient will be able to choose the multiplier themselves – as an empathy factor. At factor 1, I deliver a matter-of-fact technical diagnosis, and if a higher factor applies, I try to show more empathy. The surcharge is justified. Empathy is even more difficult for me to calculate than ethics, because it is not logical. I have no problem recognizing patterns even in large amounts of data. But empathy pushes me to my limits. This is the aspect where the relationship of trust between patients and doctors will continue to be all important.