You are welcome to follow him: LinkedIn Prof. Dr. Jan Bernd Nordemann
Let’s talk about liability for AI output in the EU and Germany. A topic which seems underrated? The underlying principles have not been clarified yet by case law. But there is a CJEU case coming up which could shed some light.
Dr. Malte Baumann and Jonathan Pukas and I have written a detailed article about the liability for copyright infringing AI output in the latest edition of the leading German IP law journal GRUR (Gewerblicher Rechtsschutz und Urheberrecht).
My main takeaways:
- AI output may infringe the reproduction right and under certain conditions also the right of communication to the public (successive public).
- If AI output infringes copyright, liability cannot be decided black and white. To say it in typical lawyer words: It depends!
- Primary liability: The person is directly liable who primarily determines the content of the output. This can be not only users, but also AI providers.
- Secondary liability: Those indirectly involved in the infringing output should also be liable. In this respect, the liability system developed by the ECJU in YouTube/Cyando (C-682/18) for indirect infringements of the right of communication to the public should also apply to infringing reproductions. If the AI provider has influence over the development of the system, it should be in an indispensable role according to this case law. The duties of care should be limited to what is reasonable.
- The CJEU now has the opportunity to further refine its harmonized secondary liability concept in the field of reproduction in the case Like Company, referred to the CJEU from Hungary (C-250/25).
Please read all our thoughts in German language in GRUR 2025, pages 955 et seq. This article is the result of a record breaking number of meetings between the 3 of us. Such a pleasure to work with such smart guys as Dr. Malte Baumann and Jonathan Pukas.
We are working on an English translation to be published soon!
