Overview
- A Munich court found that OpenAI’s language models and ChatGPT responses infringed copyrights covering the lyrics of nine German songwriters.
- Judges said an indirect perception of a work suffices to constitute reproduction under CJEU jurisprudence, covering both training use and outputs.
- The court granted most of Gema’s claims but left the method for calculating damages unresolved.
- OpenAI had not issued an immediate response; the company previously argued models do not store individual works and that users drive outputs, positions the court rejected on the facts.
- Gema brought the case in November 2024, and legal and industry groups say the first decision of its kind in Europe could shape rights enforcement and regulation across the EU.