Carlsen sues OpenAI: ChatGPT produces complete print templates for NEINhorn copies
The publisher is taking OpenAI to the Munich Regional Court alongside Marc-Uwe Kling and Astrid Henn. The core of the claim is not imitation of style, but fidelity of detail.

Illustration · AI-generated (AI IN LIFE)
At a glance
- Filed 19 August 2026 at the Munich Regional Court I against OpenAI Ireland Ltd.
- Plaintiffs: Carlsen publishing house, author Marc-Uwe Kling, illustrator Astrid Henn
- Claim: ChatGPT generates NEINhorn stories and near-identical illustrations
- Per the publisher, outputs include full print templates with cover, imprint, invented ISBN and publisher logo
- Legal argument: unlawful training and persistent "memorisation" of the works inside the model
On 19 August, a lawsuit against OpenAI Ireland Ltd. was filed at the Munich Regional Court I. The plaintiffs are the publisher Carlsen, author Marc-Uwe Kling and illustrator Astrid Henn — the creators of the German children's book series "Das NEINhorn".
The claim: on request, ChatGPT generates complete stories featuring the protected characters, and according to the publisher the resulting illustrations are barely distinguishable from the originals. The system even delivers complete print-ready templates for picture books — including cover design, imprint, an invented ISBN and the publisher's logo.
Legally, Carlsen is aiming at a specific point. From the fidelity of those outputs, the publisher infers that the original works were used unlawfully in training and persist inside the model as "memorisation", retrievable at any time through ChatGPT. The question is therefore less whether a model may imitate a style, and more whether specific protected works can be shown to sit inside it.
Author Marc-Uwe Kling names the asymmetry that carries the case publicly: anyone who illegally downloads even a single film, book or audiobook faces harsh penalties, while OpenAI helps itself to art and culture worldwide and assumes it will get away with it.
The case joins a growing set of copyright proceedings against OpenAI, but it is unusually precise in one respect: the focus is not a vague allegation about style, but a reproducible artefact — a finished, confusable print template of a protected book.
FAQ
What is the legal argument?
Carlsen contends that the highly faithful outputs prove the original works entered training unlawfully and persist as memorisation inside the model, retrievable via ChatGPT.
Who is suing?
The Carlsen publishing house together with author Marc-Uwe Kling and illustrator Astrid Henn, filed at the Munich Regional Court I against OpenAI Ireland Ltd.
What makes this case notable?
It centres on a concrete, reproducible artefact rather than a general claim about style: a complete print template including an invented ISBN and the publisher's logo.


