Spotify’s ongoing war against AI slop has now targeted tens of millions of tracks.
The music streaming giant announced Tuesday that it has removed more than 50 million AI slop and spam tracks from users’ recommendations so far this year. That’s on top of the more than 75 million spammy tracks the company says it has removed completely from its platform for serious or repeated violations of its rules.
The sheer number of tracks being purged underscores just how big of a problem AI slop and spam have become for music streaming services and the broader industry.
Spotify said it took a closer look earlier this year at tracks listeners were skipping, ignoring, or telling the company they didn’t want to hear. It found that those tracks were consistently identified as slop or spam. So, over the past six months, Spotify has made several changes to its recommendation system to weed out unwanted tracks and give music made by human artists a boost.
The company says no single signal can catch all AI slop or spam. Instead, it looks at a combination of factors, including listener engagement and feedback, upload patterns, metadata, and third-party AI detection tools.
“Using these signals together, we can identify likely slop or spam with more confidence,” Spotify said in a blog post. “We also hunt for new tactics, including mass uploads, duplicates, SEO manipulation, spammy playlist networks, and artificially short tracks. As spammers change tactics, we’ll keep updating our systems.”
Spotify says it has seen a noticeable drop in listener complaints about slop and spam appearing in recommendations since making the changes.
The announcement comes as the music industry continues to grapple with the rise of AI music-generation tools like Suno and Udio.
For instance, Universal Music Group and Sony Music Entertainment are still battling Suno in court, accusing the startup of using copyrighted recordings without permission to train its AI models. Warner Music Group, which was initially part of the lawsuit, settled with Suno last year and reached a licensing agreement allowing participating artists and songwriters to opt into approving the use of their music and likeness in AI-generated songs.
Even Suno has taken steps to crack down on the misuse of songs generated on its platform. In an August blog post, Suno co-founder and CEO Mikey Shulman announced plans to introduce audio watermarking and fingerprinting technology to make it easier to identify and trace songs created with Suno. The company has also introduced download limits intended to make it harder for users to mass-distribute AI-generated songs to streaming platforms.
These measures are primarily target schemes in which users generate hundreds and sometimes even thousands of AI songs, upload them to streaming services, and then use bots to repeatedly play the tracks to collect fraudulent royalties.
Spotify has also taken steps to address these schemes. In 2024, the company introduced a 1,000-stream annual threshold for tracks to qualify for recorded-music royalties, making it harder for scammers to make money by mass-uploading songs. More recently, Spotify has started labeling AI-generated artist personas on its platform and has verified more than 300,000 human artists, which account for over 99% of what listeners actually search for on the platform.
