Spotify starts tagging AI artists next month, and the labels have teeth

Spotify has an AI artist with nearly 168,700 followers on its platform, and until mid-September nothing on that profile will tell you a human didn’t record the vocals.

The artist is Enlly Blue. The profile photo looks like a heavily airbrushed shot of Marilyn Monroe. Blue’s biggest track, “Through My Soul,” debuted at #44 on Billboard’s “Emerging Artists” chart in October and has been streamed well over 29 million times, which works out to more than $145,000 from that song alone. Blue is a project run by a songwriter named Thong Viet, according to Billboard.

On Tuesday, Spotify said it’s going to start flagging profiles like that one.

What the label actually says

Starting in mid-September, Spotify will apply an “AI Persona” tag to artist profiles built around algorithmically generated identities. The company’s framing is unusually blunt about why.

“Listeners have been clear in telling us that they don’t like seeing an artist profile that seems human, only to find out that the persona is AI-generated,” Spotify wrote in its announcement. The label “signals to listeners that an artist’s identity may be AI-generated and does not represent a real person.”

Note the wording there. May be. Spotify isn’t claiming forensic certainty, and that hedge is going to matter the first time a real musician with an unusual aesthetic gets swept up.

The part that costs money

Here’s the piece that separates this from the usual disclosure theater: labeled AI Personas drop out of Spotify’s algorithmic recommendations entirely, unless you already follow the account.

That’s the whole business model for a profile like Enlly Blue. Twenty-nine million streams don’t come from people typing a name into search. They come from playlists and autoplay. Cut the algorithmic pipe and the economics of running a fictional pop star get considerably worse.

The tags show up in artist banners and “About” sections, in search results, and next to individual tracks in playlists. So there’s no clean surface left where an AI persona reads as human.

Confess now or get caught later

The humans behind these projects can self-disclose through Spotify for Artists. A voluntary labeling system along similar lines was proposed last month by a group of organizations representing musicians and record labels.

But Spotify will also run its own review of suspicious profiles that stay quiet, starting “with artists who have met defined audience thresholds, so the labels cover the vast majority of artist profiles that users visit.”

Spotify didn’t say what that threshold is. That’s the weak spot. A vague follower count is exactly the kind of line that gets reverse-engineered by anyone motivated to stay just underneath it, and there’s real money in staying underneath it.

The enforcement wrinkle is smarter than the threshold. Click a label and you’ll see whether the creator disclosed voluntarily or whether Spotify had to figure it out. Getting caught leaves a permanent mark on the profile. That’s a reputational cost applied to the one thing an AI persona is selling, which is the persona.

They tried this already, with a smaller ask

Spotify rolled out “AI credits” earlier this year, letting artists disclose where the technology touched a track, lyrics included.

“With AI credits, we have seen that people are eager to take the opportunity to be transparent and share how they used AI as part of their song’s creative process,” Spotify wrote Tuesday. “We’re hoping for a similar level of self disclosure for AI Personas.”

Hoping is doing a lot of work in that sentence. Admitting you used a model to help with a bridge is a very different act from admitting the singer on your profile doesn’t exist. One is a production note. The other is the product.

Why this landed now

The U.S. has no meaningful regulation around deepfakes, and streaming services absorbed the consequences. AI artists flooded in because nothing stopped them.

Musicians are split on all of this, and not evenly. Plenty will tell you it’s a plague on the industry, trained on copyrighted work nobody paid for, producing soulless filler. Others will experiment with it without handing the whole process to a machine.

And for the small group profiting off streams from artists who were never born, it’s been a windfall.

Watch what happens to Enlly Blue’s follower count after mid-September. That number, more than anything Spotify says in a blog post, will tell you whether a label attached to a profile is enough to change how people listen.