AI: Computers as intelligent aids in medicine
November 12, 2018
Everybody makes mistakes. But in some work environments, like medicine, mistakes can be deadly. That's why more and more medical personnel are turning to artificial intelligence (AI) to help reduce the rate of error.
Although, many experienced doctors are skeptical about using AI in medicine, researchers around the globe are working on new ways to apply it. The options are diverse — and in some cases rather peculiar.
Read more: Artificial intelligence, or the end of the world as we know it
Smelling diseases
Human breath contains numerous chemicals that can be helpful in the diagnosis of different diseases. But a human's nose is not sensitive enough to register and properly analyze these hundreds of chemical compounds. That is why British researchers have developed an AI-driven technology that can sniff out danger.
It works like this: Various sensors smell the patient's breath. The computer then analyzes the composition of the air and displays the measured chemicals as a 3D-graphic. Certain molecules in the breath can, for example, indicate that the patient has cancer.
These sensors not only help with diagnosis, they also store and learn from all the analyzed data from previous examinations. That means the more it's used, the more precise and effective the technology becomes over time.
Read more: AI to gobble up fewer jobs than previously thought, says OECD
Measuring pain
"Where does it hurt?" It's a common and seemingly straightforward question but one that's not always easy to answer.
Children may have difficulty in articulating their subjective pain experience, as do people with dementia, for instance.
Artificial intelligence could help through automatic pain recognition technology. Special sensors measure the body's reaction as the patient is subjected to pain stimuli. In the next step, the pain experience and level are computed. To come to a conclusive result, the AI takes various reactions into account, including breathing and blood circulation and in the skin and muscles.
Read more: The robots are coming, and they might have to pay tax
Diagnosing genetic diseases
Facial recognition can be used to determine a person's health, including diagnosing rare genetic diseases in children.
The AI-driven technology carries out a highly detailed face scan. Then a computer compares the facial features to those from a database, which includes information about various genetic diseases. Several rare defects can be diagnosed by the shape of the head or the position of a child's eyes.
Early risk recognition
Diagnostic mistakes can be fatal. But the hope is that AI will help lower the failure rate. Even though, it cannot fully substitute for an experienced doctor, it is of great value in certain fields. The technology has, for example, proven to be especially successful in diagnosing skin cancer.
One study showed that an AI-driven machine — which was trained with thousands of pictures to recognize a dangerous kind of skin cancer — was more successful than dermatologists in detecting melanoma.
Read more: Like, yesterday: Drugs and dementia discovered early
AI is also being used more frequently in gastroenterology. The technology is more effective in the diagnosis of polyps in the intestine and in determining whether they need to be removed.
A digital psychotherapist
Getting an appointment with a psychotherapist requires a lot of patience. In Germany, the average waiting time is up to six months, making it impossible for people to get help quickly. But AI is now being used to bridge the gap until a human therapist is available.
The technology is based on cognitive behavioral theory. Patients who show symptoms of depression can communicate with an AI-driven chatbot. The bot asks patients about their feelings and offers insights on how to deal with problems. It is available for conversations 24-hours a day. One study with American college students showed a decline in depression and anxiety following regular interaction with the chatbot.