Summary
- A German court has determined that Suno violated copyright by utilizing music from GEMA's catalog without proper licensing.
- The ruling mandates that AI companies secure licenses for both the training of AI models and the production of music.
- Suno has expressed disagreement with the court's decision and is contemplating an appeal.
The Munich Regional Court has ruled against Suno, an AI music startup, in a copyright lawsuit initiated by the music rights organization GEMA.
As reported by Variety, the court found that Suno infringed copyright laws by training its AI models with music from GEMA's catalog and by reproducing copyrighted works without authorization. This ruling establishes that AI companies must obtain licenses for the commercial use of GEMA's music, which includes both AI model training and music generation.
The lawsuit specifically highlighted six songs, including "Daddy Cool," "Rasputin," "Forever Young," and "Mambo No. 5," with GEMA alleging that Suno trained its AI on these copyrighted tracks without consent.
In response, Suno stated that its platform was designed to assist users in creating original music rather than duplicating existing tracks. The company believes the ruling stems from a misunderstanding of its technology and is currently exploring legal avenues.
“Our tools empower users to craft new songs, whether they are established artists, product developers, or casual music fans,” Suno commented. “Since our inception, we have trained our models to generate original music, not replicate existing songs, and we have integrated safeguards into our platform. We disagree with today’s decision, which mischaracterizes how Suno’s technology functions and its application under U.S. law, and are reviewing all options, including an appeal.”
This ruling follows a previous copyright victory for GEMA against OpenAI, where a Munich court determined that ChatGPT had unlawfully reproduced song lyrics.
Additionally, Suno is currently under increased legal scrutiny regarding the training of its AI models. Recent leaks revealed that the company's training data comprised over 113,000 hours from YouTube Music, 62,000 hours from Pond5, and 12,000 hours from Deezer, which supports industry claims that Suno trained its AI using copyrighted material without authorization, allegations that the company continues to contest in court.