On Tuesday, November 11, 2025, a regional court in Munich ruled that OpenAI’s chatbot ChatGPT violated German copyright law by using unlicensed song lyrics to train its language model. This landmark decision stems from a case brought by GEMA, the German music rights society representing composers, lyricists and publishers.
The court found that ChatGPT was trained on protected content from nine German songs, including best-selling musician Herbert Groenemeyer’s hits "Maenner" and "Bochum." Presiding judge Elke Schwager ordered OpenAI to pay damages for the unauthorized use of these lyrics, though the exact figure has not been disclosed.
GEMA’s lawsuit highlights a growing clash between rights holders and AI developers over data scraping practices. Artists worldwide have raised concerns that large language models are ingesting copyrighted text without permission or compensation, potentially undermining creative industries.
For AI companies and startups, the Munich ruling raises urgent questions about legal compliance in model training. As OpenAI and others expand their datasets, they may need to secure licenses or develop new content-filtering strategies to avoid similar disputes.
Legal experts say the decision could reverberate across Europe and beyond, prompting regulators and courts to scrutinize how AI systems use copyrighted material. It also underscores the value of clear guidelines and fair licensing frameworks to balance innovation with creators’ rights.
OpenAI has not yet announced whether it will appeal the ruling. In the meantime, the case sets a precedent for both AI developers and rights holders as they navigate the rapidly evolving intersection of technology, law and creativity.
Reference(s):
Training ChatGPT on lyrics infringed copyright, German court rules
cgtn.com
