German performance rights organisation GEMA has introduced PLAI, a licensing service that offers a dataset of nearly 180,000 music recordings cleared for use in training artificial intelligence systems.
The launch follows nearly two years of development since GEMA first outlined plans for a licensing framework covering generative AI. The Berlin-based society said the dataset spans more than 60 genres and includes comprehensive metadata, with all rights cleared for AI training purposes.
Initial rights holders and compensation
Participating music providers at launch include Bailer Music Publishing, Earmotion Audio Creation, Intervox Production Music, Music Sculptor and Sonoton Music, all based in Germany. GEMA stated that participating rights holders will receive proportionate and appropriate compensation, and the platform remains open to additional publishers, collecting societies, authors and other rights providers.
Tailored data and first customer
The service is designed to give AI developers a single source for licensed music, with data packages that can be customised by genre, mood and other parameters. GEMA said the offering aims to provide greater legal certainty for companies building AI tools.
AI-powered music transcription company Klangio is the first announced customer. Klangio head Sebastian Murgul said access to legally cleared training data is critical for the business.
“Our customers and we ourselves are part of the music industry,” Murgul said. “That is why fairness and legal certainty are a major issue for us. With the GEMA music dataset, we have found the solution we have been waiting for: legally fully clarified training data where all rights are really covered.”
He added that the arrangement provides the security to develop AI tools on a solid foundation and demonstrates that fair remuneration and innovative AI development can coexist.
Legal context and the Suno case
GEMA director of broadcasting and online Thomas Theune said PLAI simplifies access to licensed music while creating new income potential for creators.
“Instead of complex rights clearances, AI tool providers now receive high-quality music data including the necessary rights from a single source,” Theune said. “This creates legal certainty for developers and new revenue potential for music creators.”
Theune added that the service shows AI training and fair remuneration of creatives are not contradictory, provided both sides engage responsibly.
The launch coincides with GEMA’s ongoing legal action against AI music company Suno over the use of copyrighted works in training. A ruling in that case is expected on 31 July. GEMA has previously won a court dispute involving OpenAI.
GEMA general counsel Kai Welp said the Suno outcome could have broader implications for licensing generative AI systems. He stressed the need for clarity that using creative works for training requires remuneration, regardless of where the training occurs or whether original works are stored in the systems.
“For the licensing of generative AI providers, we hope for a positive decision on the matter against Suno on 31 July,” Welp said. “It must be made clear that the use of creative works must be remunerated regardless of the location of the training and also when the original works are stored in the systems. This requires the courts, but above all the political framework.”
GEMA’s PLAI model seeks to position licensed datasets as an alternative to potentially unlicensed uses of copyrighted music, offering AI developers a structured route to obtain training material.
