Every week we read about new clinical pictures, that can be diagnosed faster and now with greater certainty than humans, using an algorithm. While doctors have to inform themselves about current studies and guidelines in addition to their clinical work and bureaucracy is accounting for an increasing share of working time, an algorithm can fall back on enormous computing capacities. In some areas of expertise, such as oncology or radiology, algorithms are already a step ahead of doctors. They will probably be of great help in medicine in the future. However, we must not forget that so-called neural networks only detect correlations, but never prove causalities. Furthermore, we must pay attention to which companies process our data so that monopolies are not promoted. With this commentary article we would like to point out the possibilities and possible risks of artificial intelligence in medicine and ask further questions regarding the future training of medical doctors the influence of politics and what role physicians will play in the future.