Our partner Prof. Dr. Jan Bernd Nordmann regularly posts on copyright topics on LinkedIn.
You are welcome to follow him: LinkedIn Prof. Dr. Jan Bernd Nordemann
AI output is increasingly a focus of copyright law. When does an infringement occur? Who is liable? These are all good questions. I attempted to answer them in a presentation at the GRUR – Deutsche Vereinigung für Gewerblichen Rechtsschutz und Urheberrecht e.V. Berlin Group. Here are my main takeaways:
- Infringement in AI output needs to be assessed according to the CJEU criteria in “Mio and Konektra.” Yes, if copyright-protected elements are found in the AI output.
- For liability, a distinction must be made between direct and indirect causation. If the AI system directly causes the infringement, it is liable.
- The distinction (between direct and indirect causation) should be made based on who primarily determines the AI output. According to the Munich Regional Court in “GEMA/OpenAI,” the AI is liable if the infringement can be repeatedly generated through easy prompting.
- In cases of indirect causation by the AI system, the standards developed in the CJEU case law “YouTube/Cyando” should apply to both communication to the public and reproduction.
- More details? See Dr. Malte Baumann/ Dr. Jonathan Pukas/ Jan Bernd Nordemann in German in GRUR 2025, 955 – and in English on SSRN: https://lnkd.in/d8n8tHpi
