Spotify Pulled More Than 50 Million AI Tracks From Its Recommendations

Spotify has put a number on how much AI-made filler it is keeping away from listeners. In a one-year update on its AI protections published October 20, the company said it has taken more than 50 million "slop and spam" tracks out of recommendations this year alone, according to Spotify's own post.

That is separate from more than 75 million spammy tracks it says it has removed from the platform entirely for serious or repeated rule violations.

The distinction matters. Tracks pulled from recommendations have not necessarily been deleted. They are kept out of personalized feeds and editorial playlists, and Spotify frames its approach as anti-spam and anti-deception rather than a ban on AI music. A few details show how it works:

• No single signal flags a track. Spotify combines listener engagement, upload patterns, metadata and third-party AI detection tools, per Billboard's summary
• Thousands of artist profiles now carry an AI Persona badge, and those artists are excluded from editorial and algorithmic recommendations by default unless a listener follows them
• Spotify says tracks with strong signs of being AI-generated are ten times more likely to fall below its 1,000-stream royalty threshold, and that listener complaints about AI spam have decreased markedly

Spotify also claims it is the only major streaming service that labels AI-generated artist identities. Rivals are moving too: Deezer says it receives nearly 90,000 fully AI-generated tracks a day and keeps them out of recommendations, while Tidal withholds royalties from tracks it identifies as fully AI-generated.

The open question is accuracy, since all of these numbers come from the platforms' own detection and none has been independently audited, and a system that relies on several signals can wrongly flag human artists as easily as it catches mass-produced spam. Spotify itself says its work is not finished, which is probably the most reliable line in the announcement.