Spotify says AI-flagged tracks are 10x more likely to miss its royalty threshold

What happened

Spotify published an update on its AI rules on October 20. Three facts in it matter if you make or commission music.

First, the AI Credits system the company launched in April is now taking tens of thousands of AI-use disclosures a day, according to Spotify. Artists can flag AI involvement in vocals, instruments, lyrics or production. Listeners see it in Song Credits on mobile.

Second, AI Personas — artist profiles that are themselves AI — are excluded from editorial and algorithmic recommendations by default. They surface only if a listener follows them.

Third, and this is the part with money attached: Spotify says tracks carrying signals of a high probability of being AI-generated are 10 times more likely to stay below the platform's 1,000-stream royalty threshold. Below that line, a track earns nothing.

The question stopped being whether you can upload AI music and became how you disclose it if you want to be found.

What this is not

It is not a ban. Spotify is not removing AI-assisted tracks, and disclosing AI use does not get you delisted. A human artist who used an AI tool on a vocal stack is not in the same bucket as a bot profile uploading 400 tracks a week.

It is not a visible label in the player. The disclosure lives in Song Credits, which most listeners never open. If you were expecting a YouTube-style banner, that is not what shipped.

It is not clearance. Spotify's credit field says how a track was made. It says nothing about whether you have the rights to use that track in a video, a podcast, or an ad. Those are separate questions and nobody answered them this week.

And it is not a published detection method. Spotify has not disclosed which signals trigger the AI-probability flag, what share of royalties is affected, or whether there is an appeals process. Those gaps could not be confirmed. Treat the 10x figure as the company's own number, not an audited one.

Why the threshold does the real work

The 1,000-stream minimum was not built for AI. It was built to stop fractions-of-a-cent payouts on dead catalogue. But it works as an AI filter without Spotify having to defend a content judgment, and that is why it is the more consequential lever here.

If you are an agency or publisher who ran the math on generating a large library of background tracks and collecting a trickle per track, the math changed. Volume was the whole strategy. Volume is what the threshold eats.

The open question is direction. Spotify could use disclosure to give transparently human-led, AI-assisted work better discovery. Or it could use detection mainly to suppress catalogue flooding and leave disclosed work no better off. Nothing in the October 20 post settles it.

This is also not happening in isolation. Trade coverage this month has tracked parallel tightening on synthetic uploads at YouTube, TikTok and Instagram. The direction across platforms is the same: disclosure required, distribution conditional.

What to do about it

1. Start an AI-origin log for every track you release: tool, model, prompts or source material, human performers, editing steps. One row per track.
2. Put AI credits in the release metadata at delivery, not in marketing copy after the fact.
3. Ask your distributor two questions in writing: does it pass Spotify AI Credits through, and does it support a detection dispute if a track gets flagged.
4. Pull any revenue forecast that assumes income from a large volume of generated tracks. Rebuild it with the 1,000-stream floor applied per track.
5. Separate human-led releases from AI-heavy experiments in your reporting, so one does not hide the other's performance.
6. Keep project files, stems and session records. If a classification is disputed, that is your only evidence.
7. For video and podcast use, keep sourcing your beds from licensed libraries. A Spotify credit is not a license.