Artificial intelligence – how far can it go in dermatology?

The late Professor John McCarthy, known as the ‘Father of Artificial Intelligence’ who coined its name in 1955 and won the Kyoto Prize for his work in 1988, defined artificial intelligence (AI) as ‘the science and engineering of making intelligent machines, especially intelligent computer programs’.1


In 2016, we are now on the cusp of a major healthcare revolution leveraging AI capabilities. Electronic medical records provide an ever-growing wealth of data that lend themselves to analysis by complex algorithms that can identify patterns that humans simply cannot. The groundwork done by tech giants (e.g. Google, Facebook, Microsoft, Apple and Amazon) in developing AI to deliver tailored search results has set the stage for more accurate and/or timely disease detection, more effective disease prevention, more efficient patient management and more personalised treatment using AI.2


In the US, healthcare reform together with the availability of more powerful and more affordable computers has led in part to the integration of new technologies into the healthcare setting in order to improve outcomes and reduce costs. Modernizing Medicine is a Florida-based company that with physicians has developed a cloud-based repository of medical information and insights generated from millions of patient visits and associated treatments, with the aim of increasing efficiencies and improving treatment and business outcomes. Their speciality-specific applications include one for dermatology. Modernizing Medicine provides recommendations based on physicians’ most popular actions for a given scenario, such as drug prescribing or lab test ordering; it also gives data on patient outcomes that can be cross-referenced with that of the latest clinical research by sending a request that is handled by IBM’s AI supercomputer Watson. Collaboration between Modernizing Medicine and IBM has included integrating Watson into a software package for dermatologists to speed up medical coding and note taking. In addition, hypothesis-generating capabilities are set to aid physicians in the treatment of skin cancer and other dermatological diseases.3


Although AI applications are still in the very early stages of development in the healthcare sector, their potential is far-reaching. In some centres in the US, home-grown systems have been developed to provide pop-up notifications to alert physicians when a particular drug may not work with the patient at hand based on genetic traits or to predict what treatments may be required in the future. It has been reported that doctors may heed the computer’s advice in up to two-thirds of cases. On a broad scale, access to high-quality data and incompatibilities between different electronic medical record databases are current challenges. However, Eric Schmidt, Chairman of Alphabet Inc. (parent company of Google), recently cited an AI-based dermatology company as an example of the perfect start-up company. Using crowdsourced input from dermatologists, Schmidt imagined a highly accurate AI diagnosis tool could be built through machine learning and deployed via smartphone technology.4


An obvious concern is the potential for AI to encroach on physicians’ practice. Quite where the line will be drawn between making recommendations and clinical decisions remains to be seen. One thing that seems certain at this stage is, you can’t expect to purely rely on a computer to tell you what to do, and medical judgement that includes a conversation with the patient is best left to the treating physician.

Dominique du Crest


  1. McCarthy J. What is artificial intelligence: basic questions. Nov 2007. Available at: Accessed July 2016.
  2. Hernandez D. Artificial intelligence is now telling doctors how to treat you. Kaiser Health News. June 2, 2014. Available at: Accessed July 2016.
  3. Neil Ungerleider. IBM’s Watson is ready to see you – in your dermatologist’s office. May 16, 2014. Available at: Accessed July 2016.
  4. Eric Schmidt’s idea of a perfect startup company. News Republic. June 17, 2016. Available at: Accessed July 2016.