OpenAI has lodged an appeal against a Munich court ruling that found the company had breached German copyright law in a dispute brought by the music rights society GEMA over song lyrics used in training ChatGPT, according to Reuters and a statement from the company. The move sets up what is likely to become one of the most closely watched AI and copyright cases in Europe.

GEMA alleges that OpenAI used lyrics from protected German songs without permission and that its chatbot can reproduce them word for word when prompted. The collecting society says the case raises a basic question for the generative AI industry: whether companies may train systems on copyrighted works without licences and then profit from the results.

OpenAI argues the German court misunderstood how language models work. The company says its systems do not store training data as searchable copies, but instead learn statistical patterns from large datasets and generate new output in response to user requests. The Munich court rejected that defence, saying the ability to reproduce lyrics verbatim reflected memorisation rather than mere statistical pattern matching, and it ordered OpenAI to stop using the material without a licence and to pay damages.

Legal specialists have described the ruling as potentially significant beyond Germany because there is still limited settled case law on the issue across Europe. Reuters reported that a similar GEMA case is also pending in Munich against the AI music generator Suno, involving well-known songs including "Atemlos" and "Daddy Cool", underlining how far the dispute may spread as courts test how copyright rules apply to generative AI.

Source Reference Map

Inspired by headline at: [1]

Sources by paragraph:

Source: Noah Wire Services