Kennt die künstliche Intelligenz Humor? KI-Generierung deutscher und ägyptisch-arabischer Witze. Eine kognitiv-semantische Studie

Jokes constitute a text type that is strongly shaped by culture, whose success depends largely on structural patterns, shifts in meaning, and the disruption of habitual interpretive frameworks. the present study examines AI-generated German and Egyptian Arabic jokes from a cognitive-semantic perspective on both the structural and content levels. The aim is to determine the extent to which artificial intelligence is capable of reproducing culture-dependent joke mechanisms and triggering an intended laughter effect. Central to the analysis are the cognitive-semantic processes underlying joke generation, as well as the role of culturally shared knowledge about the societies represented in the jokes. The analysis draws on key approaches in cognitive semantics, particularly with regard to meaning construction and the formation of the punchline. In addition, the study investigates which constitutive prerequisites of jokes can be identified in AI-generated texts. Another focus lies on the question of whether, and in what way, the cultural background of the joke characters influences the structure and effect of the jokes. Through a contrastive comparison of German and Egyptian Arabic examples, the study contributes to an understanding of the cultural adaptability of AI and provides insight into the extent to which AI possesses human-like cognitive abilities in the domain of humorous communication.