According to 404 Media, a hacker has breached the company database of Suno, an online AI music generation tool. The hack is said to reveal multiple references to the training data it scraped from YouTube Music, Deezer, Genius, and others.
Suno has previously admitted that its tool was trained on “essentially all music files of reasonable quality that are accessible on the open internet” as part of court proceedings back in 2024.
The company has argued that it’s allowed to train on copyrighted works under fair use protections, while the Recording Industry of America has accused it of unlawfully scraping tracks from YouTube specifically, among other complaints.
And looking at the details, it appears YouTube Music was indeed a primary target. A file entitled “youtube_music” reportedly contains a note that over two million music clips were scraped, while another dataset file notes a figure of “113,879 hours of youtube_music.” Alongside “17,615 hours of genius_hq,” “12,287 hours of Deezer”, and “152,162 hours of ytm_tagged.”