When artificial intelligence is discussed today, public debates usually oscillate between two extremes. On the one hand, there are everyday office applications that have long become established through language models and automated data processing. Writing emails, summarizing long text documents, or automatically creating presentations are now part of the standard repertoire of modern administrative structures. On the other hand, future debates often paint a picture of fully autonomous systems that completely replace human decisions, independently control complex processes, or watch over industrial workflows as detached control authorities.