Suno, an AI music generation startup, has admitted in a court filing that it trained its AI model using copyrighted songs but claimed this was legal under the fair-use doctrine. The admission follows a lawsuit filed by the Recording Industry Association of America (RIAA) on June 24, alleging that Suno and another startup, Udio, used copyrighted music without permission. Suno’s CEO, Mikey Shulman, stated in a blog post that the company trains its models on high-quality music available on the open internet, including copyrighted materials owned by major record labels. He argued that this process is akin to a musician learning to create new music by listening to existing songs, asserting, “Learning is not infringing. It never has been, and it is not now.”
The RIAA responded by accusing Suno of industrial-scale infringement, arguing that their use of copyrighted material does not qualify as fair use. The association emphasized that such actions undermine artists’ ability to earn a living, stating, “Their vision of the ‘future of music’ is apparently one in which fans will no longer enjoy music by their favorite artists because those artists can no longer earn a living.”
Suno’s defense hinges on the claim that using data from the open internet, including copyrighted songs, is a fair and legal way to train AI models, comparable to human learning processes. The outcome of this legal battle could have significant implications for the future of AI in the music industry and the interpretation of fair use in the context of AI training.
Read the full story here.