Suno hack reveals millions of songs from YouTube, Genius, and Deezer were used to train its AI model

Suno, the AI-generated music platform that’s currently being sued by both Universal Music Group and Sony Music, and recently acquired the concert discovery app Songkick, has been hacked. And the results revealed that “millions” of songs from YouTube, Genius, and the French streaming platform Deezer have been used to train Suno’s AI models, as have multiple stock music libraries. 

Suno has previously admitted to training its models on copyrighted material, saying in one of their many legal proceedings that “essentially all music films of reasonable quality that are accessible on the open internet” were acceptable AI training material under fair use laws. The Recording Industry Association of America previously accused Suno of ripping songs directly from YouTube, an accusation that the hack confirms.

Multiple artists including SZA and Kenneth Blume have railed against the use of their music in AI model training. The former said she was “certain” that even some of her unreleased songs were used to train AI models, and argued that AI exploitation disproportionately affects Black artists and their work. 

 One might assume that the Suno hacker, who has not yet revealed their identity, shares a similar stance on AI and its consequences for musicians. But in an interview with 404 media, the hacker denied having any vendetta against AI music or Suno in particular, and that their motivation was random: “I like to hack anything and everything.”

Keep scrolling for more great stories.