A Munich court delivered a landmark verdict against AI music maker Suno. On July 31, 2026, the regional court ruled Suno infringed copyright by training its AI on protected music without permission.
Suno’s AI was trained on over two million scraped songs, including six tracks represented by GEMA, Germany’s music collecting society. GEMA represents more than 95,000 creators and over 2 million rights holders worldwide. The court found Suno memorised and reproduced those songs during training.
GEMA argued that using copyrighted music without a license violates rights holders’ protections. The court agreed, stating Suno had no right to process or reproduce works by GEMA-represented artists. The ruling requires Suno to disclose revenue tied to the infringement for damage calculations.
This decision is immediately enforceable under German law. Tobias Holzmüller, CEO of GEMA, said their goal isn’t to eliminate Suno but to force fair licensing negotiations, something Suno had refused so far.
Suno’s platform generates about 7 million tracks daily, with 75,000 ending up on streaming services. Their AI creates songs from text prompts in under a minute. This scale makes the ruling’s impact profound, potentially disrupting how AI music is trained and monetized.
The case reflects wider legal battles over AI-generated content. Major labels like Sony, Universal, and Warner called on July 29, 2026, to ban AI “slop” tracks from music charts, arguing only human-made songs should count. Streaming companies face pressure to add AI transparency labels.
The European Union also plans regulations requiring AI companies to watermark or label AI-generated content. Meanwhile, artists are pushing back. Lawsuits from illustrators and authors claim AI companies misused their work without consent or payment.
Notable plaintiffs include Sarah Andersen, Karla Ortiz, and Kelly McKernan, who sued Stability, Midjourney, DeviantArt, and Runway AI. Author Kirk Wallace Johnson found his books used to train AI models without permission. Meta faced a lawsuit for using authors’ books to train its Llama AI.
Legal outcomes vary. Judge William Alsup ruled training on legally acquired books for AI was transformative fair use. But other cases ended in hefty settlements, like the $1.5 billion payout by Anthropic after artists complained of copyright misuse.
Google’s Lyria AI music engine also faces lawsuits. Lead plaintiff Sam Kogon argues AI is devaluing musicians’ work and giving it away free, threatening artists’ livelihoods. Google counters that YouTube’s terms give them broad content rights. Lawyer Krystle Delgado disagrees, saying few expect to surrender remake rights by uploading videos.
This Munich ruling sets a new European precedent. It signals courts may hold AI companies accountable for training data choices. The message is clear: AI cannot freely mine protected creative works without permission or risk legal consequences.
Based on
- A German court says AI music maker Suno broke copyright, a first for Europe — thenextweb.com
- German lawsuit tests copyright limits for AI music training — dw.com
- Major Labels Call For AI Slop Songs To Be Banned From Entering Charts Around The World — engadget.com
- EU begins crackdown on AI risks | AP News — apnews.com
- Artists are lawyering up against AI slop, and some are even winning | The Verge — theverge.com
