A hack of AI music platform Suno has revealed that the company scraped song and lyric data from major streaming and reference sites, including YouTube, Deezer and Genius, to train its generative music technology. The breach, first reported by 404 Media after an unnamed hacker shared data with the outlet, also exposed Stripe payment details belonging to hundreds of thousands of Suno users, raising fresh privacy concerns for the platform's customers. The revelations add to mounting scrutiny of Suno, which is already the subject of multiple lawsuits from artists and record labels over alleged copyright infringement.

Leaked data showed millions of songs and lyrics from leading streaming services had been used to train Suno's models, alongside stock libraries such as Pond5, Jamendo, Freesound and the International Music Score Library Project, plus podcasts gathered via RSS feeds. Suno said the breach occurred in November 2025 but involved only "outdated source code" no longer in use, maintaining that its models are trained on "publicly available" files and metadata found online. The company, the largest AI music generator on the market, previously denied using copyrighted recordings without permission but has since shifted to a fair-use defence, which rights holders continue to dispute. By November last year Suno was reportedly generating a volume of music equivalent to Spotify's entire catalogue every month.