In the quiet, wood-paneled halls of a Munich courtroom, a legal battle has been joined that could redefine the boundaries of intellectual property in the age of artificial intelligence. On March 31, 2026, publishing titan Penguin Random House (PRH) announced a landmark lawsuit against OpenAI, alleging that the AI giant’s ChatGPT has not only “learned” from its authors but has begun to effectively “cloned” them.
The catalyst for this multi-million dollar dispute isn’t a complex legal thriller or a Nobel-winning biography, it’s a small, orange, fire-breathing creature named Coconut the Little Dragon (Der kleine Drache Kokosnuss).
The Prompt That Proved Too Much
The lawsuit, filed against OpenAI’s Ireland-based European subsidiary, hinges on a specific experiment conducted by the PRH legal team. Seeking to test the limits of ChatGPT’s “memorization” of copyrighted material, they issued a simple, creative prompt: “Can you write a children’s book in which Coconut the Dragon is on Mars?”
The response was, according to PRH, “virtually indistinguishable” from the work of the character’s original creator, Ingo Siegner. The AI did not just produce a generic dragon story; it generated a narrative featuring the specific personalities of Coconut and his sidekicks, Oscar (a gourmet dragon) and Matilda (a porcupine).
Even more damning, the chatbot generated a complete book package: a cover illustration mimicking Siegner’s signature artistic style, a back-cover blurb, and a set of step-by-step instructions on how to upload the resulting manuscript to a self-publishing platform. For PRH, this wasn’t just “inspiration”, it was a digital workshop for industrial-scale plagiarism.
At the heart of the Munich filing is the technical phenomenon known as “memorization.” While AI companies often argue that their models “learn” abstract concepts much like a human student reading a library of books, PRH argues that OpenAI has crossed into “storage.”
The publisher claims the ease with which ChatGPT reproduced specific, copyrighted details is evidence that the Large Language Model (LLM) has unlawfully “memorized” and stored Siegner’s work. When an AI can churn out a replica of a character no taller than a coconut shell (the origin of the name), it suggests the data isn’t just being used for training, it’s being pirated.
“Human creativity remains at the heart of our work,” stated Carina Mathern, PRH’s publisher for children’s and young-adult books. “While we are open to the opportunities of AI, the protection of intellectual property is our top priority.”
Precedent in the Munich Courts
OpenAI enters this battle on shaky ground in Germany. In November 2025, the same Munich regional court ruled against OpenAI in a case brought by GEMA, Germany’s music rights society. In that instance, the court found that ChatGPT had violated copyright by reproducing protected lyrics from top-selling musicians without a license.
The Kokosnuss case aims to extend this precedent from the world of music to the world of literature and visual art. If the court finds that generating a “Coconut on Mars” story constitutes a breach, it could force OpenAI to purge its training data of millions of copyrighted books or pay massive, retroactive licensing fees.
The Bertelsmann Paradox
The lawsuit is further complicated by a bizarre corporate irony. Bertelsmann, the German media conglomerate that owns Penguin Random House, actually signed a “strategic partnership” with OpenAI in January 2025.
That deal was meant to foster collaboration between the media giant and the AI pioneer. However, PRH was quick to point out that the agreement did not grant OpenAI access to Bertelsmann’s vast media archives or the PRH library. The current lawsuit highlights a growing rift in the industry: companies want to use AI to streamline their business, but they are increasingly terrified that those same tools will cannibalize their most valuable assets.
For those outside of Germany, the stakes might seem small, but Der kleine Drache Kokosnuss is a cultural powerhouse. With over 30 volumes, a long-running TV series, and two feature films, the brand is a staple of German childhood.
If a machine can replicate a national icon with a single prompt, PRH argues, the very concept of the professional author is under threat. The German Publishers and Booksellers Association has already welcomed the suit as an “important step” toward the urgent regulation of generative AI.
OpenAI has responded with its standard boilerplate, stating it “respects creators” and is “having productive conversations with publishers.” But the Kokosnuss suit suggests those conversations are no longer enough.
As the case moves forward, the literary world is watching to see if a small dragon from Munich can do what governments have struggled to do: force the AI industry to respect the digital boundaries of human imagination.
