Munich court finds Suno infringed GEMA music copyrights
Developed over time first seen 2 months ago
A Munich court has ruled that generative music platform Suno infringed copyrights held by Germany’s collecting society GEMA by training on, storing and reproducing protected songs. The decision matters because it supports the principle that AI companies must obtain licences for copyrighted music used both to train models and to generate outputs, reinforcing rights holders’ position in European AI litigation.
GEMA’s case focused on six songs, including “Atemlos”, “Rasputin” and “Mambo No. 5”, which it said Suno could closely reproduce using basic prompts. GEMA represents around 100,000 members and said Suno had used their works without permission; Suno had previously stated that its training data included most reasonably available music online. Suno rejected the ruling’s reasoning, maintaining that its models were designed to create new songs rather than reproduce existing ones.
- Munich court finds Suno infringed GEMA music copyrights.
- Ruling supports licensing for AI training and outputs.
- Suno disputes the decision and its interpretation.
New here? Start with this
Generative music tools use artificial intelligence to create songs from written prompts, such as a style, mood or subject. They are trained on large collections of existing recordings and musical works, which has raised questions about whether permission is needed when copyrighted music is used.
GEMA is Germany’s main organisation for collecting royalties for songwriters, composers and music publishers. It licenses music on their behalf and pays them when their work is used; its members include both individual creators and rights-owning companies.
The dispute is part of a wider international argument between technology firms and the music industry over AI training data and outputs that resemble existing songs. Courts’ decisions can affect how AI music services obtain licences, pay rights holders and operate in different countries.
Both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
Supporters of the ruling argue that creators should retain meaningful control over how their work is used, including as material for commercial AI systems. If a platform can reproduce recognisable songs from simple prompts, they say this is more than learning general musical ideas: it risks substituting for licensed music and undermining composers’ income. Requiring licences for training and outputs would make AI development accountable and ensure that the economic benefits are shared with the people whose work made it possible.
The case against
Critics argue that treating training as a licence-dependent use of every work could make music AI prohibitively difficult to develop, particularly for smaller innovators. They contend that models analyse patterns across large bodies of material rather than keeping a usable copy of each song, and that legal rules should distinguish this from direct reproduction. A broad licensing requirement, they say, may entrench established rights holders and limit new creative tools, while narrower safeguards against outputs that are substantially similar to particular works could better balance innovation and protection.
More coverage
AI Cybersecurity Elections Entertainment Music Politics Technology
Read the full article at the source →
Originally published by Variety as “Suno Loses Landmark AI Lawsuit to German Performing Rights Society GEMA”.