Germany Fines OpenAI for Illegally Using Song Lyrics to Train ChatGPT
11 days ago
- #AI
- #legal
- #copyright
- A Munich court ruled that OpenAI infringed German copyright law by using song lyrics in its training data and ordered compensation to the writers.
- The court found that both storing the lyrics internally and their appearance in ChatGPT outputs violated copyright.
- GEMA, representing over 100,000 composers and publishers, filed the claim in November 2024 over nine unlicensed German songs.
- OpenAI argued its systems adjust settings based on patterns, not copying specific works, and users are responsible for outputs—both arguments were rejected.
- OpenAI disagreed with the ruling, stating it affects only a limited set of lyrics and doesn't impact daily users in Germany.
- GEMA called the judgment the first major European ruling on copyrighted material in AI training, setting a precedent for creative works.
- The German Journalists’ Association hailed the verdict as a milestone for copyright law, potentially influencing U.S. lawsuits against OpenAI.