Suno AI Trained on Leaked Deezer, YouTube, Pond5 Data

Suno AI, a music generation platform, was trained using thousands of hours of audio data sourced from platforms like Deezer, YouTube, and Pond5, according to leaked source code. The revelation, detailed in a report by The Verge, suggests that Suno's training dataset included substantial copyrighted material without explicit permission from the rights holders.

The leaked code, which originated from a former Suno employee, outlines a process where audio was scraped from these platforms. This includes music from Deezer's catalog, user-uploaded content from YouTube, and stock audio from Pond5. The sheer volume of data used raises significant questions about copyright infringement and the ethical implications of AI training practices.

While Suno has not officially commented on the leaked code, the findings point to a potential violation of intellectual property rights. The use of copyrighted music for training AI models has become a contentious issue across the AI industry, with artists and rights holders expressing concerns about fair compensation and unauthorized use of their work.

This incident highlights the ongoing debate surrounding the legality and ethics of AI training data. As AI models become more sophisticated in generating creative content, the provenance and legality of the data used to build them will continue to be a critical point of scrutiny for both regulators and the creative industries.