Jason Moran reported an unauthorised EP, For You, on his Spotify profile. The Guardian documented his complaint on April 11.
An artist’s name should be one of the least mysterious things about a release. If the music does not belong to the person whose page you are visiting, the streaming service has handed you the wrong record in a very convincing sleeve.
The wider issue brings together three different problems: music being attached to the wrong profile, synthetic tracks being uploaded at scale, and automated plays being used to seek royalties. They can overlap. Treating them as the same thing, though, makes it harder to see what needs fixing.
A Wrong Credit Is a Problem Even Without AI
Spotify acknowledges that music can land on the wrong artist’s page through metadata errors, artists sharing a name or deliberate misattribution. Its Artist Profile Protection feature is being tested as an optional, limited beta to give artists more control over releases attached to their identity.
That distinction matters. An incorrect credit does not by itself establish that a song was generated by AI, that the account was hacked or that someone successfully collected money. Each of those claims needs its own evidence.
For listeners, the immediate failure is simpler: the page no longer reliably answers “is this by the musician I came to hear?” For artists, correcting it creates work they should not have to do. Creative freedom ought to include the freedom not to release an album.
Deezer’s Numbers Show Upload Pressure, Not a Majority of Listening
In its April 20 announcement, Deezer says it receives nearly 75,000 fully AI-generated tracks a day, about 44% of daily uploads. It reports that these tracks account for only 1–3% of total streams.
Deezer also says up to 85% of streams of fully AI-generated tracks in 2025 were detected as fraudulent, and that detected fraudulent streams are excluded from royalty payments. The company describes tagging AI-generated music and removing it from recommendations.
These are Deezer’s findings about its own service and detection system. They are not a measurement of Spotify’s catalogue, and they do not mean that nearly half of everything people listen to is synthetic.
The denominators tell different stories. A large share of new uploads can coexist with a small share of listening. And attempted manipulation is not the same as a successful payout when the platform identifies and excludes the plays.
There Is a Documented Fraud Case—With a Different Date
The financial risk is not hypothetical. The US Department of Justice announced on March 19, 2026 that Michael Smith pleaded guilty to conspiracy to commit wire fraud in a scheme involving AI-generated music and automated streaming.
According to the announcement, hundreds of thousands of generated songs received billions of bot plays, producing more than $8 million in fraudulent royalties. The scheme spread activity across many tracks to avoid making any single song’s numbers look too conspicuous.
That is a concrete example of generated content and artificial listening being combined for financial gain. It does not establish that every disputed jazz release belongs to the same operation, or that all AI-assisted music is fraudulent. The offence is not simply that a computer was involved in making a sound.
Profile Protection Helps, but Coverage Matters
An artist-review layer addresses one part of the problem: whether a release belongs under a particular name. Its usefulness depends on access, clear notifications and a process that artists or their authorised teams can actually manage.
A limited beta is not universal protection. Nor is rejecting a misattributed release on one service necessarily a correction everywhere that distributor delivered it. This is why communication between the artist, label, distributor and platform matters.
There are trade-offs in proposed fixes, too. Automatically freezing every artist who has not released music recently could obstruct legitimate archival releases or a return after a long break. Quiet periods are not evidence that an artist’s identity should be retired by a timer.
The aim should be dependable attribution and effective handling of disputes. A dashboard can help, but making musicians perform endless administrative solos is a rather joyless addition to the job.
What Listeners and Artists Can Check
- Look for the artist’s own announcement. Their website, established social accounts or label can help confirm an unexpected release.
- Check the credits and label. An unfamiliar entry is a reason to investigate, not automatic proof of fraud.
- Do not diagnose AI from artwork alone. A strange cover or a change of musical style cannot settle who made a recording.
- Use the official reporting route. Artists should involve their authorised team, label or distributor when a release is incorrectly attached to them, and check whether the profile-protection beta is available.
None of this requires assuming jazz audiences are unusually gullible or that older musicians cannot speak for themselves. The issue is the reliability of the catalogue, not the age or taste of the person pressing play.
Music discovery works better when surprise means finding something wonderful, rather than discovering that the name on the page is wrong. The cat tolerates a wide range of jazz, but would also like the correct artist credited before walking across the keyboard.
🐾 Curiosity looks good on you. Explore the Goodies, or find our illustrated books on Amazon.
