The serious flaw in Spotify’s plan to distinguish humans from AI

The serious flaw in Spotify’s plan to distinguish humans from AI

Spotify’s new method of distinguishing AI-generated music from that made by real people is laudable… but sadly flawed, says Tom Service

© MARIA CORTE MAIDAGAN


Spotify's green tick identifies the artist as human

Whoops! It’s happening again: this column tends once again towards the subject of AI and the disastrous era of recording sound that is upon us, out there in the regions marked ‘mega-corporate streamers’ and ‘artificially generated bands who rack up millions of plays’ – musical places we all fear to tread.

But wait! There’s hope, because Spotify has launched its ‘Verified by Spotify’ green tick, to prove to anyone listening that what you’re experiencing when you stream music by the artist known as Taylor Swift or Timothy Ridout has been created by an authentic human being – or as Spotify says, by a musician who shows ‘defined standards demonstrating authenticity’. Which actually isn’t quite the same as being a musical human, but given that Spotify recentlty removed 75 million AI-generated tracks, let’s take the win. 

Spotify’s criteria of musical humanity are the following: that artists have social media accounts associated with their profiles, as well as merchandise and concert dates, the sorts of things that live human musicians and composers get up to. So what could possibly go wrong?

The flaw in the Spotify verification system

Well, quite a lot. Firstly, it’s a wild fact of our digital lives that it’s humans who need verification, not the bots. Why not a red cross of despair for an AI-generated track and leave the humans without needing to be verified? It’s therefore a working assumption that anthing not verified by Spotify’s tests is an AI-bot making music with the sole purpose of racking up plays from listeners, who themselves may or may not be human.

The second problem is how Spotify’s tests favour musicians with already established careers, enabling them to have the social media presences, merchandise and concert calendars to prove they’re human. That means those who are starting out can’t be recognised as ‘authentic’, because apart from uploading their songs to Spotify, what’s to prove they’re the real thing? In giving the pretence of looking after human musicians, Spotify’s verification may have the opposite effect, lumping the unverified bots together with the unverifiable early-career musicians, who can’t prove their authenticity – the ultimate digital Catch-22.

No ticks for Beamish or Lachenmann

And there’s more. You too can play a grimly fascinating game to find out who’s human and who’s not, according Spotify. Ed Sheeran? Ticked. Dua Lipa? Another reassuring green tick. But composer Sally Beamish, despite her recent more-than-a-million streamed track Gerropaedie? No tick at the time of writing. Helmut Lachenmann? No tick either. 

Having met them both, I know that Beamish and Lachenmann are real human beings with more ‘authenticity’ in their little fingers than entire genres out there on Spotify. I’ll say it again in this column: classical music from Beamish to Beethoven is human-made music to its core and offers musical culture its best defence against the ravages of AI. Go on Spotify, give them all a green tick!  

Footer banner
This website is owned and published by Our Media Ltd. www.ourmedia.co.uk
© Our Media 2026