Deezer Takes Action Against AI-Generated Music
Deezer announced on Friday its decision to label albums containing AI-generated tracks as part of an initiative to combat streaming fraud. According to the company, approximately 18% of music uploaded daily, totaling over 20,000 tracks, is now fully AI-generated. Despite most of these tracks not gaining widespread popularity, around 70% of their streams are considered fake, intended to generate royalties fraudulently.
In response to this issue, Deezer has implemented clear labeling for AI-generated tracks, excluding them from editorial playlists and algorithm-based recommendations. Additionally, measures have been taken to filter out fraudulent streams from royalty payments. This new approach aims to assist listeners in distinguishing between human-created music and AI-generated content.
Growing Trend of AI-Only Songs
Deezer reports that AI-only songs currently represent just 0.5% of all streams on its platform, but this percentage is rapidly increasing. CEO Alexis Lanternier emphasized the company’s commitment to transparency and user trust, stating that AI is neither inherently good nor bad. He stressed the importance of a responsible and transparent approach in the context of ongoing debates about copyright law and AI development.
Technological Innovations to Address the Issue
In December 2024, Deezer applied for two patents related to its AI Detection technology, which focuses on identifying unique signatures to differentiate between synthetic and authentic content. This initiative coincides with discussions between major record labels such as Universal Music Group, Warner Music Group, and Sony Music Entertainment, and AI startups Udio and Suno. The record companies have filed lawsuits against these startups for copyright infringement, with potential licensing agreements aiming to resolve the legal disputes.
