Book Market: Can AI Author Literature?
An expert at LMU Munich discusses AI's role in literature and copyright. Current AI models struggle to create suspense, but challenge our understanding of creativity.
Professor Julian Schröter at Ludwig-Maximilians-Universität München (LMU) has analyzed the role of artificial intelligence (AI) in the literary field. According to him, current language models, as of mid-2026, still struggle to create compelling suspense. Schröter attributes this to AI's limited "theory of mind," its inability to adequately anticipate and understand reader expectations and thoughts.
The complex art of building suspense in literature proves to be a significant challenge for AI, even though it can recognize simple deceptive intentions. Schröter highlights that AI cannot yet construct narratives where unexpected events ultimately appear necessary and plausible. This reveals the profound complexity of suspense storytelling, an aspect previously underestimated by literary critics.
While AI's use in literary production is considered a potential tool, the realm of "high literature" is not context-free. Schröter points out that it is a complex social construct of attributions and conditions that define what is considered high literature. The human author remains a central element of this definition. He cites author and Nobel laureate Olga Tokarczuk, who faced criticism for using AI but later clarified it was for research purposes.
Copyright issues arise when AI generates content that closely resembles existing works. Particularly in illustrated children's books, such as "Kokusnuss the Dragon," AI has produced continuations that strongly mimic human-created styles. Schröter suggests this challenges the concept of plagiarism and necessitates new regulatory solutions to protect creative work. The European AI Act aims to address these concerns.