Spotify has removed more than 75 million AI-generated tracks from its platform over the past year. The company says the tracks were mainly spam and were removed to protect legitimate artists and songwriters from losing royalties.
The move does not mean that Spotify is banning AI music. The platform still allows music created with the help of artificial intelligence, as long as the creators have the necessary rights and have made a genuine creative contribution.
According to reports from the Los Angeles Times and Rolling Stone, Spotify’s main target is not AI itself, but fraud, mass uploads and tracks created only to generate streaming revenue.
The rise of AI music spam
AI tools have made it possible to create and upload thousands of songs in a very short time. Some users have taken advantage of this by flooding streaming platforms with simple, repetitive tracks.
These uploads can include short ambient songs, piano loops or sounds such as rain and white noise. In some cases, the tracks are uploaded in large numbers and promoted through artificial streams or bots.
Spotify executives have referred to this type of content as “AI slop”. The term describes music created from simple prompts, uploaded in bulk and designed mainly to take advantage of the way streaming royalties are distributed.
The problem is not completely new. In 2023, Spotify removed tens of thousands of songs connected to the AI music platform Boomy after detecting artificial streaming activity. As AI tools have become cheaper and easier to use, however, the amount of low-quality and spam content has grown significantly.
How Spotify is fighting back
Spotify has introduced several measures to identify and remove suspicious tracks:
- A ban on unauthorized voice cloning: AI-generated copies of recognizable artists’ voices can be removed if they are used without permission.
- Machine-learning spam filters: Spotify looks for unusual behavior, including mass uploads, duplicate tracks and artificial streaming.
- Improved metadata standards: Labels and creators can indicate whether a track was fully generated by AI or only used AI in certain parts of the production, such as vocals, mixing or mastering.
- Protection against data scraping: Spotify is also working to stop AI companies from collecting large amounts of music from its catalogue to train their models.
- A dedicated monitoring team: The company is tracking new spam techniques and other forms of streaming fraud as they appear.
Spotify’s policy team told Rolling Stone that spammers should not be earning royalties that should go to professional artists and songwriters.
AI music is still allowed
Spotify has made it clear that the use of AI is not automatically a reason for removing a song.
AI-assisted mixing and mastering are allowed. Authorized remixes and covers are also permitted, as long as the creators have the necessary rights.
The line is crossed when users upload thousands of short, almost identical tracks with the clear intention of generating streams. For example, an account that floods Spotify with hundreds of similar “relaxing rain sounds” tracks could be flagged and removed.
Spotify is also working with record labels to help establish which AI-created music can be legally released and monetized on the platform.
Concerns from AI music creators
Not everyone is fully comfortable with Spotify’s approach. Some AI music creators worry that aggressive filters could accidentally remove legitimate experimental projects.
There are also concerns that labeling music as AI-generated could create a negative impression, even when a human artist has played an important role in the creative process.
Organizations such as IFPI and RIAA are supporting voluntary labeling systems that would allow listeners to see whether a track was fully generated by AI or created with AI assistance.
The battle is far from over
Spotify’s removal of 75 million tracks shows how serious the problem of AI-generated spam has become. However, the company’s efforts are unlikely to end here.
As Spotify improves its detection systems, spam operations may find new ways to avoid being caught. Some may start adding small human elements to AI-generated tracks in an attempt to make them appear more legitimate.
The bigger question is whether streaming platforms will eventually offer listeners separate filters for human-made and AI-assisted music. For now, Spotify is drawing a clear distinction: AI can be part of the creative process, but it should not be used to manipulate the streaming economy.
