Hacker Reveals Suno AI's Vast Data Scraping Practices

In November 2025, AI music generation platform Suno experienced a limited security incident that was quickly contained. The breach reportedly exposed outdated source code, including instructions for scraping training data from various platforms. This data purportedly includes millions of music clips and tens of thousands of hours of audio scraped from services like YouTube Music, Genius, and Deezer, as well as stock music libraries.

A Suno spokesperson confirmed the incident, stating that the exposed source code was no longer in use and that no sensitive personal information was compromised. The company also asserted that the hack did not reveal anything new about their training practices. Suno claims its models are built around 'Original Creation, By Design,' intentionally avoiding artist names as training metadata and implementing filters to prevent replication of existing works.

Despite these claims, Suno faces lawsuits from major music labels alleging its AI models can generate outputs that directly imitate artists' styles and melodies. Suno's defense rests on the argument that its training data, comprising publicly available music files from the open internet, constitutes fair use. The legal standing of this fair use defense remains an open question, with past rulings being divided.