[Artificial intelligence in medicine : Foundations, potential, structural limits and current developments]

Artificial intelligence (AI) is widely regarded as a key technology for the further development of medical care. Its performance, however, depends less on algorithmic innovations than on the quality, availability and interoperability of medical data. This article outlines the foundations of medical AI, current fields of application, with a particular focus on oncology as well as the relevant ethical, legal and quality assurance frameworks. Central to the discussion is the hypothesis that AI can only realize its full potential if it is built upon FAIR (findable, accessible, interoperable, reusable), semantically interoperable and standardized data and infrastructures. The AI is not an optional add-on but a necessary infrastructural prerequisite for modern medicine.