Suno caught using millions of YouTube songs to train AI
AI music platform Suno is in hot water over its training data. Hacked internal code shows it grabbed millions of songs and lyrics from YouTube Music, Deezer, Genius, and more to build its AI model.
The crawler pulled over 113,000 hours of content from YouTube Music, 17,000 hours of lyrics from Genius, and 12,000 hours of songs from Deezer. It also took from Pond5 and Jamendo. Suno previously used vague terms like 'public files' to describe its training methods.
Worse, the hacker also accessed Suno's customer database, exposing emails, phone numbers, and partial credit card data. Suno decided not to notify users, calling it a 'limited security incident' and claiming no legal obligation to alert them.
Comments
Loading comments…