Time for ‘Doctor AI’ to take the lead

  • Share
    Two clicks for more data privacy: click here to activate the button and send your recommendation. Data will be transfered as soon as the activation occurs.
  • Print
  • Read out

Medicine is the area that holds out the best hopes for the use of artificial intelligence (AI). Humankind’s dream of coming up with cures for diseases such as cancer or Alzheimer’s seems within grasp. The use of algorithms to analyze health data may provide entirely new insights into diseases and cures. But from a very personal perspective, from the patients’ viewpoint, AI is also bringing about change. We talked about this aspect with the well-known physician and comedian Dr. Eckart von Hirschhausen.

AI detects skin cancer more reliably than dermatologists

AI is very good at recognizing images. The ability to detect patterns can greatly assist the treating physician to diagnose skin cancer, for instance. Researchers from the University of Heidelberg have discovered that AI is better than humans at providing an early diagnosis of dangerous melanomas.

Trained with 100,000 images of dangerous or benign skin changes, AI competed with 58 dermatologists from various countries. The algorithm correctly identified 95 percent of the dangerous skin changes, well ahead of the dermatologists. The dermatologists in turn were more effective at spotting benign moles – an equally important quality as it prevents unnecessary operations. This is a good example of the fact that humans and machines work best as a team. Physicians can save time by getting the AI to initially analyze the image and then checking the result. This reduces their workload, giving them more time for the patient: ”Time with the patient is what counts. If digital solutions help me spend more time looking after patients, talking to them, explaining things to them, then that’s great!,” Eckart von Hirschhausen believes.

AI also helps with premature infant deaths in intensive care. The evaluation of large amounts of data on the vital functions of premature infants has revealed that all vital functions of such babies suddenly became extremely stable about 24 hours before the outbreak of a life-threatening infection. AI was the first to identify this typical, dangerous progression. In this way, physicians can now act promptly and continue to monitor the signs despite the apparent stabilization of the babies’ condition.

The patient as physician

Eckart von Hirschhausen is, however, skeptical of the ability to perform a “self-diagnosis” on the net. An internet search for illnesses, symptoms, or treatment options, for instance throws up an infinite number of hits. But which pages really offer serious content and medically sound advice? “What worries me is these black-box algorithms deciding what is relevant and what not. If I enter serious health issues into the Google search box, the engine comes up with myriad hits. But what isn’t clear is whether the engine takes me to pages containing evidence-based, scientifically sound information, whether I’m dealing with maverick views, whether the hits are advertising-driven, or whether someone is trying to sell me certain drugs. All that must be transparent. And it isn’t transparent at present.”

So to use technology, you need to define a framework – and that necessitates rules and reliable recommendations. Find out about other aspects in the video interview with Eckart von Hirschhausen. He talks about how the internet has not made the world smarter but rather dumber. About patient power and about their own responsibility, and, of course, about humor, which he sees as the last area where humans have the edge over the smartest machines.

Enjoy watching the interview!

Eckhardt von Hirschhausen

AI in medicine

Eckart von Hirschhausen on artificial intelligence, patients and humour.

FAQ