German court hears GEMA copyright case against OpenAI over ChatGPT training
A German music rights body sued OpenAI in November 2024, alleging that ChatGPT was trained on copyrighted German song lyrics without permission. The Munich Regional Court confirmed the chatbot can reproduce the lyrics faithfully, and will rule on the case on November 11, 2025.", "event_date":"2024-11", <parameter name="facts">[ { "claim": "GEMA sued OpenAI alleging ChatGPT was trained on lyrics of nine German songs without permission.", "source_url": "https://economictimes.indiatimes.com/tech/artificial-intelligence/german-music-body-sues-openai-alleging-copyright-breaches/articleshow/124216313.cms" }, { "claim": "The Munich Regional Court confirmed that when given simple prompts, ChatGPT reproduced large parts of the song lyrics faithfully.", "source_url": "https://economictimes.indiatimes.com/tech/artificial-intelligence/german-music-body-sues-openai-alleging-copyright-breaches/articleshow/124216313.cms" }, { "claim": "OpenAI argues its large language models do not store or copy specific data but rather reflect what they have learned from training data.", "source_url": "https://economictimes.indiatimes.com/tech/artificial-intelligence/german-music-body-sues-openai-alleging-copyright-breaches/articleshow/124216313.cms" }, { "claim": "The Munich court will rule on the case on November 11, 2025.", "source_url": "https://economictimes.indiatimes.com/tech/artificial-intelligence/german-music-body-sues-openai-alleging-copyright-breaches/articleshow/124216313.cms" } ]
What happened
GEMA, a German music rights organisation representing over 100,000 composers, songwriters and publishers, sued OpenAI on November 2024 alleging that ChatGPT was trained on lyrics of nine German songs without permission.
The Munich Regional Court confirmed that when given simple prompts, ChatGPT reproduced large parts of the song lyrics faithfully, establishing that the lyrics are memorised in OpenAI's large language model.
Reported impact
GEMA alleges the training constitutes copyright infringement and a systematic unauthorised use of its repertoire, seeking unspecified damages.
Index analysis
OpenAI disputes the allegations, arguing that its large language models do not store or copy specific data but rather reflect what they have learned from training data. The company contends that users who generate output from the chatbot are responsible for that output, not OpenAI.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.