Suno snatched millions of songs from YouTube, Genius, and Deezer

Editor's take

Suno's AI music generation model was reportedly trained on a vast dataset of copyrighted songs and lyrics extensively scraped from platforms like YouTube Music, Deezer, and Genius. This revelation directly challenges the legality of Suno's training data, raising significant questions about intellectual property rights and fair use in the context of AI development. The implications extend to artists whose work may have been used without explicit consent or compensation, and to the broader music industry grappling with the disruptive potential of generative AI.

The exposure of Suno's data sources necessitates a closer examination of how AI models are being built and the ethical boundaries of data acquisition. Future developments will likely focus on legal challenges from copyright holders and potential regulatory responses aimed at ensuring transparency and fairness in AI training practices. It will be crucial to observe how Suno and its competitors address these data provenance issues and whether they adopt more ethically sourced or licensed datasets to mitigate future risks.