Künstliche Intelligenz in der Medizin

Machine learning technologies based on a big data approach are seen as the key to personalized medicine. From image recognition software to chatbots, from decision support systems to social robots, these technologies offer a wide range of possible uses in medicine. The particular challenge is that these applications of artificial intelligence (AI) are not just passive tools but artificial agents, i.e., non-human actors who make autonomous or semi-autonomous decisions and take actions. In addition, these technologies change the epistemic practices of doctors, i.e., the way doctors “read” patients. Both aspects lead to fundamental transformations of practices, relationships and environments in medicine. This article explains the ethical aspects of this transformation. Whether medicine will change for the better or worse is not determined by technology; rather, it is our crucial task to define clear goals for development, implementation and use of the technology. The principle should be: automate processes not decisions; delegate time-consuming activities instead of replacing doctors. In this way, the potential of AI to enable best practice and promote patient well-being can be unleashed. These considerations are based on the fundamental results of the author’s monograph Ethics of Medical AI.