Contact
Your law firm Geißler Legal.
address
Eupener Straße 59
D-50933 Cologne
Opening hours
Mon-Sat: 10:00 am – 1:00 pm
Mon-Fri: 2:00 pm – 8:00 pm
and by telephone appointment
Geissler Legal - Legal Advisor for Compliance & Commercial Law
GEMA vs. Open AI - Copyright protection vs. Data mining
Compliance-focused legal advice for your business activities - personal, cross-border, cost-efficient.
Contents
1. The ruling of the Munich Regional Court I - A milestone in copyright law
In a ruling dated November 11, 2025, the Munich Regional Court I (Case No. 42 O 14139/24) prohibited OpenAI – the operator of ChatGPT – from, among other things, using certain protected song lyrics:
- to store in its systems,
- to be used for content generation,
- and to reproduce publicly,
unless a valid license exists. This significantly strengthens the artists' rights enshrined in copyright law – with considerable consequences, as this article will explain in more detail.
In this specific case, the issue concerned nine well-known song lyrics, including those by Herbert Grönemeyer, Helene Fischer, and Reinhard Mey, which were reproduced by ChatGPT without the copyright holders' consent. The court clarified that reproducing such content—even in part—violates the copyright of reproduction (§ 16 German Copyright Act) and the right of making works available to the public (§ 19a German Copyright Act). A clear statement, but one that, from a purely legal perspective, is mandatory and therefore only logical.
This constitutes a clear violation of copyright law (UrhG) – regardless of whether the use is purely algorithmic or human-controlled.
2. Scope far beyond the individual case - protection for all intellectual property rights holders
As a lawyer specializing in copyright and media law, I see this ruling as having far-reaching implications for all creators of intellectual works of every kind – not just for prominent artists.
Copyright protection applies to every personal intellectual creation as defined in Section 2 of the German Copyright Act (UrhG) – including photographers, copywriters, musicians, designers, and authors. This is because, particularly in the context of so-called data mining – the automated analysis and extraction of copyrighted content – works are increasingly being fed into AI systems, processed, and output without consent.
Many copyright holders are unaware that simply saving or processing their works using AI models can constitute copyright infringement – even without the specific content being published. The crucial factor is the technical capture and use without a license.
3. Copyright claims - injunction, disclosure, damages
Anyone who discovers that their works are being used or reproduced without a license by AI systems like ChatGPT or similar services can take legal action. The key claims are:
- Claim for injunctive relief (§ 97 para. 1 Copyright Act) – to terminate the use
- Right to information (§ 101 Copyright Act) – to determine the type and intensity of use
- Claim for damages (§ 97 para. 2 Copyright Act) – regularly calculated according to the license analogy
The license analogy allows copyright holders to demand appropriate compensation for unauthorized use – regardless of whether a license agreement actually existed. This calculation is based on standard market rates, which can be determined, for example, via GEMA tariffs or image fee schedules.
4. Consequences for platforms, AI providers and data mining projects
For providers of AI technologies, platform operators and companies in the field of machine learning, the ruling of the Munich Regional Court I becomes a key compliance risk.
The judges clarified: Using, storing, and distributing copyrighted content without a license is illegal – regardless of whether the material was publicly available on the internet. Invoking the limitations on copyright for text and data mining (Sections 44b et seq. of the German Copyright Act) does not automatically apply, especially when it involves the output of copyrighted works.
Conclusion: Anyone who accesses copyrighted content in their AI or data mining projects must have legally compliant licensing models – otherwise, they risk injunctions, claims for damages and legal action, including from abroad.
5. Recommendations for authors, musicians and photographers
Anyone who works creatively – whether as a musician, photographer, writer or publisher – should now check whether their works:
- captured and processed by AI models,
- made publicly available or issued,
- or were fed into training databases.
As a lawyer specializing in copyright and media law, I support copyright holders in this endeavor.
- to enforce information claims,
- to assert license claims,
- to calculate damages based on the license analogy,
- and to create legally sound license agreements in order to avoid future legal infringements.
6. Conclusion - Copyright in the Age of AI and Data Mining
The ruling of the Munich Regional Court I is a clear sign in favor of copyright law:
Artificial intelligence does not operate in a legal vacuum.
Those who create have rights.
Anyone using it needs a license.
For creators, this means that independent control over intellectual property remains essential even in the age of AI. For companies and platforms, this means that data mining requires legal clarity.
Make an inquiry now
We will be happy to advise you comprehensively and personally on your concerns.
legal area
No posts found
English 
German 
Italian