Spotify does not label AI music, and users are building their own tools to detect it
The website 404 Media describes how the amount of unlabeled AI music on Spotify is growing, prompting independent developers to create the detection tools SoullessMusic and SlopTracker, while Spotify did not respond to questions.
The website 404 Media (investigative technology journalism) described the case of a singer performing under the name Slime Dot, an R&B artist from Las Vegas with more than 100 000 monthly listeners on Spotify. In an interview with the magazine GQ, Slime Dot denied being an AI avatar; GQ initially published the interview without reservations and only later added a note to the text stating that the character was AI-generated. Unlike YouTube, Spotify does not label AI-generated music in any way, which, according to the article, contributes to similar misunderstandings.
The lack of labeling has led to the creation of independent tools for detecting AI music. The website SoullessMusic.com, created by Graeme Fulton, maintains a database of “AI artists hidden on Spotify” and identified Slime Dot as a track almost certainly created using AI, based on an analysis of three tracks and an unusually high rate of new music releases. The SlopTracker tool allows users to upload a file or a link to a track on Spotify and assesses the probability that it is AI-generated content – for the track “Fully” by Slime Dot, it determined a 95% probability that it was created using Suno. According to the article, similar cases include releasing AI music under the name of a deceased musician to farm streams, a similar case involving a Danish jazz musician, or the channel Qajar Jazz, which has almost 30 000 monthly listeners, releases hours of Iranian jazz every week on Spotify, and is labeled in small print as AI-generated content on YouTube (unlike on Spotify). The channel Qajar Jazz did not respond to a request for comment, nor did Spotify.
For detection, Fulton uses the open models SONICS and the so-called lofcz vocoder fakeprint detector, checks track metadata, and compares artist names with Wikipedia entries; reported tracks also undergo manual review, but he says the system is not flawless and, as with SlopTracker, false positives can occur. According to an estimate by SoullessMusic, a small group of monitored AI artists collectively generates around 5.7 million dollars a year, with the most successful of them, the account mikeeysmind, alone generating approximately 1.5 million dollars a year. Last year, Spotify announced that it would introduce labeling for AI-generated music, but the author of the article found no such labels on any of the dozens of tracks checked, even though the music generation tools Udio and Suno embed an inaudible digital watermark in the audio that would make labeling easier. By contrast, according to an April statement from the company, rival service Deezer labels AI music on its platform and said that 44 % of newly uploaded music is already AI-generated.
Details, including additional figures on the financial estimates from the SlopTracker tool, can be found in the source article.
Why it matters
Without mandatory labeling of AI music on Spotify, listeners and musicians have to rely on external, imperfect detection tools, and the line between real and AI-generated artists is becoming blurred even in the media, as the case of the magazine GQ shows. For the music business, this raises the question of where some of the streaming revenue is going and how competing platforms such as Deezer address the same problem with a different strategy.
Two audiences, two different impacts
What this means
For individuals
People who listen to music on Spotify cannot be certain that an artist they follow is a real person – a suspicious track can be checked using the independent tools SoullessMusic or SlopTracker.
For a business
Streaming platform operators face the risk that unlabeled AI music diverts revenue from real artists and undermines the credibility of the catalog, while competitor Deezer pursues the opposite strategy of transparent labeling.
Risks and complianceCheck the original
Event sources
only one source so far · 1 publisher, 1 independent. We count feeds from the same owner only once.