German Court rules in favour of GEMA in Copyright case against OpenAI

A German court in Munich has ruled in favour of the country’s music rights society GEMA in a high-profile copyright case against the U.S.-based artificial intelligence company OpenAI, marking one of Europe’s first legal decisions on AI and creative rights.

The Regional Court of Munich held that OpenAI’s chatbot ChatGPT had reproduced copyrighted song lyrics without proper licensing, thereby violating copyright law. The court, presided over by Judge Elke Schwager, ordered OpenAI to pay damages to GEMA for the unauthorised use of protected material. It was ruled that song lyrics cannot be used by AI systems without obtaining a valid licence from rights holders.

GEMA —(Society for Musical Performing and Mechanical Reproduction Rights) represents over 100,000 composers, lyricists, and music publishers in Germany. Its members include well-known artists such as Herbert Grönemeyer. The society had argued that ChatGPT was capable of reproducing lyrics from copyrighted German songs without authorisation and that OpenAI’s training datasets contained protected content from its members’ repertoire.

In response, OpenAI stated that GEMA’s claims were based on a misunderstanding of how ChatGPT functions, asserting that the system does not intentionally store or reproduce specific works. Despite this, the court sided with GEMA, agreeing that training and output generation involving copyrighted material require explicit permission or licensing.

Legal experts have noted that this ruling could set a precedent for how generative AI systems are regulated across Europe, especially in the wake of the EU Artificial Intelligence Act passed in 2024. The verdict is expected to influence how AI developers disclose training data and handle copyrighted works going forward.

GEMA has long been vocal about the unauthorised use of musical works by AI systems. In 2023 and 2024, the organisation had warned global AI companies such as OpenAI, Google, and Stability AI against training their models using copyrighted content without paying royalties. The present case marks GEMA’s first major legal victory after months of discussion about AI’s use of protected cultural material.

Across Europe, similar concerns have been raised by other rights groups, including France’s SACEM (Society of Authors, Composers and Publishers of Music), which has sought a collective licensing framework for AI-generated content. The European Union’s transparency rules now require AI firms to disclose whether copyrighted material was used in model training, a move intended to safeguard creative ownership.

Globally, comparable lawsuits have been filed in the United States by artists and authors — including comedian Sarah Silverman and the Authors Guild — who alleged that OpenAI and Meta had used their works without consent. However, no final rulings have yet been delivered in those cases.

In Germany, GEMA has expressed optimism that the court’s decision will pave the way for a clear licensing system for AI training and output, ensuring that creators are fairly compensated. OpenAI and GEMA have both stated that official responses to the verdict would be released later on Tuesday.

The decision can still be appealed by OpenAI, meaning the legal debate over how artificial intelligence interacts with copyright-protected works is likely to continue in higher courts.