source: techcrunch ai: hack suggests ai music generator suno scraped youtube for training data

level: technical

a hacker breached ai music generator suno in november 2025 using a supply chain attack to steal employee credentials. the intruder accessed source code that allegedly shows how suno scraped decades of audio from youtube music, deezer, genius, stock music libraries, and podcast rss feeds. customer data including emails, phone numbers, and partial credit card numbers were also exposed. suno did not notify users about the breach, calling it a limited incident that was quickly contained.

suno has previously admitted training on publicly available music files from the open internet, claiming fair use protection. major record labels suing suno argue that circumventing youtube's anti-scraping measures violates the digital millennium copyright act and youtube's terms of service. competitor udio faces similar accusations of scraping youtube data. google, youtube's parent, is also dealing with copyright lawsuits from book publishers over training data practices.

the hack reveals the tension between ai development and copyright law. suno's alleged scraping of copyrighted audio without permission highlights the legal risks for generative ai companies. the breach also raises concerns about security practices, as customer financial data was compromised without disclosure. this incident may influence ongoing lawsuits and regulatory scrutiny of how ai models obtain training data.

why it matters: the breach exposes how ai music generators may be using copyrighted material without permission, which could reshape legal boundaries for training data and force companies to change their data collection methods.


source: techcrunch ai: hack suggests ai music generator suno scraped youtube for training data