A significant security incident has affected Suno, an AI music generation service, with a hacker reportedly breaching the company's database and revealing the extent of its data collection practices. According to reports, the breach has exposed details about the vast amounts of music data Suno has scraped from platforms like YouTube Music, Deezer, and Genius.
Suno has previously acknowledged in legal proceedings that its AI models were trained on "essentially all music files of reasonable quality that are accessible on the open internet." This statement was made as the company defends itself against accusations of unlawful data scraping, particularly from YouTube, by entities like the Recording Industry of America.
The leaked information suggests that YouTube Music was a primary source, with files indicating the scraping of over two million music clips and approximately 113,879 hours of content. Additional datasets mention significant hours of audio from Genius and Deezer, alongside "ytm_tagged" content. The company also reportedly utilized third-party proxies and tools like PodcastIndex to gather audio media, potentially including hundreds of thousands of podcasts.
A spokesperson for Suno stated that the company's AI models have been trained on publicly available music files and metadata. They also noted that in November 2025, Suno identified and contained a limited security incident involving outdated source code, asserting that no sensitive personal information was compromised.
This incident highlights ongoing debates surrounding AI training data. Legal precedents exist where AI companies have successfully argued fair use for copyrighted content used in training models. For instance, a US judge ruled in favor of Anthropic regarding copyrighted content for AI training, and Meta also won a similar case concerning training data derived from authors' books. Suno maintains that its generated outputs are distinct from original works and that it has implemented safeguards against misuse.
However, the music industry expresses considerable concern. Musicians and industry representatives have voiced apprehension about AI's role in music creation, with some distinguishing between AI as a creative assist tool and generative AI for direct output. The core issue remains the use of artists' work for training AI models without explicit consent or compensation, fueling ongoing industry-wide discussions about copyright, fair use, and the future of music creation in the age of artificial intelligence.