Suno trains AI with YouTube music without permission

Investigations revealed that Suno, a company that offers an AI music generator, collected millions of audio tracks from YouTube, Genius and Deezer and from rights holders. These songs were used as training data for its music generation models.

The case comes at a time when the creative community and copyright holders are increasingly attentive to the use of protected content in AI training, sparking debates about licensing, compensation and the limits of fair use.

For AI students and enthusiasts in Brazil, this episode highlights the importance of understanding where the data that feeds models comes from, as well as alerting to the ethical and legal responsibilities when creating or using content generation tools.

Imagine a developer creating an app that generates soundtracks for games using Suno; if the training data comes from unlicensed content, the product could face infringement claims and need to be taken off the market or renegotiated.

When evaluating generative AI tools, check whether vendors disclose the sources of their training data and whether they have proper licenses; prefer platforms that offer transparency or that use open domain or licensed data.