German court rules OpenAI's ChatGPT infringed song lyrics copyright
A Munich court ruled that OpenAI violated German copyright law by using song lyrics to train ChatGPT and allowing the chatbot to reproduce them without authorization or payment. The court ordered OpenAI to compensate GEMA, a music rights organization, for damages including unpaid royalties. OpenAI said it would appeal the decision.
What happened
On November 11, 2025, a Munich court ruled that OpenAI infringed copyright by using song lyrics to train ChatGPT and reproducing them in the chatbot's outputs without authorization.
The case was brought by GEMA, Germany's largest music rights management society, representing around 100,000 music industry professionals. GEMA alleged that OpenAI used lyrics from nine German songs to train its language models and that ChatGPT reproduced these lyrics when users requested them, without licenses or compensation to creators.
Reported impact
The court found that both the memorization of lyrics in the training data and their reproduction in ChatGPT responses constituted copyright infringement. The tribunal ordered OpenAI to indemnify GEMA for all damages, including unpaid royalties and court costs, potentially amounting to several hundred thousand euros if the decision is upheld on appeal.
Index analysis
The court applied European case law establishing that indirect perception of a work constitutes reproduction sufficient for copyright infringement. The tribunal rejected OpenAI's defense that its models do not store individual data but instead learn patterns from training sets, ruling that users did not cause the copyright violation but rather OpenAI made the song lyrics publicly accessible through ChatGPT without authorization.
Governance controls
OpenAI announced it would appeal the decision and stated that it respects the rights of creators and content holders, claiming to conduct constructive discussions with organizations worldwide. A company spokesperson noted the decision concerned only a limited set of lyrics and did not affect millions of users in Germany.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.