Spotify says it has excluded more than 50 million tracks it classifies as “AI-generated slop and spam” from recommendations so far in 2026. The tracks remain available in its catalogue but no longer appear in algorithmic or editorial recommendations.
The company announced the figure in an October 20 blog post reviewing the first year of its artificial intelligence (AI) safeguards, introduced in September 2025.
“AI-generated slop and spam have no place in your recommendations, and we have made changes to keep them out and protect the listening experience.”
Changes to recommendations
Spotify said it analysed tracks that listeners skipped, ignored or said they did not want to hear earlier this year. It found that those tracks were consistently identified as slop or spam.
The company said it made “numerous updates” to its recommendation system over the past six months to exclude that content and give greater visibility to music by “authentic human artists”. Its editors apply the same criteria and do not programme content identified as slop or spam.
Since those changes, Spotify says it has seen “a notable decline” in listener complaints about encountering that content in recommendations.
The post also cited more than 75 million spam tracks removed entirely for serious or repeated rule violations. That figure was previously disclosed in September 2025 and covered the preceding twelve months. Spotify continues to remove content that breaches its rules but has not shared an updated total.
How Spotify identifies content
Spotify combines listener engagement and feedback, upload patterns, metadata and information from several external AI detection tools to identify the material.
“AI detection technology has limitations, and it will become increasingly difficult as generative AI evolves.”
The company also monitors tactics including:
- Bulk uploads and duplicate tracks.
- Search engine optimisation (SEO) manipulation.
- Spam playlist networks.
- Artificially short tracks.
Royalties and artist safeguards
Spotify describes its 1.000-stream threshold for earning royalties, in place since 2024, as “a first line of defence for human artists”. According to its data, tracks with a high likelihood of being AI-generated are 10 times more likely to fall below that threshold.
The company also provided figures on its artist tools:
- More than 300.000 verified artists account for more than 99% of searches on the platform; more than 70% of those artists are independent.
- Artists submit “tens of thousands” of AI Credits each day.
- Artist profile protection, already being tested by thousands of musicians, will become available to all artists before the end of the year.
Other platforms introduce AI controls
In the past week, YouTube Music began labelling AI-generated music and Deezer announced plans to let listeners hide it from search results. Apple Music already uses transparency labels supplied by labels and distributors.
Spotify says it is currently the only major streaming service that flags when it believes an artist’s identity is AI-generated.
“This work is not finished. AI is advancing fast, and bad actors are adapting just as quickly,” the company said.
All figures come from Spotify and have not been independently verified. The company has not disclosed how many of the 50 million excluded tracks were still receiving plays or what proportion were identified using external tools.
