How Do You Know a Song Is AI? A Music Curator’s 4 Steps to Sniff Out AI Slop

How do you know a song is AI? A music curator shares the 4 steps they use to sniff out AI slop, from bios to their discography, before a song makes the playlist.

SELECTOR'S NOTES

Dakota Nash

11/17/20267 min read

robot and human hands reaching toward ai text
robot and human hands reaching toward ai text

I was on YouTube Music, and I put on a radio station built off "Lost in Rhythm" (Supernova Remix). I let the algorithm do its thing, and a few

songs in, one track caught my ear. You know the feeling. You stop what you're doing, check your phone, and need to know who made this.

So I tapped on the artist. And instead of finding a story, I found a wall of music. Release after release, all from 2026. I wanted to go deeper, but something felt off. I started questioning the name. Then the visual identity. Then the whole thing.

That moment changed how I think about my job as a curator. It's easier than ever to find good music right now, partly because of what AI is capable of. That's exactly why I have to be more diligent. Every song I put in a TLR Session is a promise to you that a real human made it.

So how do you know a song is AI?

Here's the exact process I use to vet every artist before they get anywhere near a Session, and why I think it matters more than ever.

Why You Can't Trust Your Ears Anymore

Here's what nobody wants to admit: the scary part isn't that AI music is bad. It's that it's good. Like, really freaking good.

Take Sienna Rose. She showed up as a neo-soul singer with smoky, jazzy vocals, and her songs took off. According to Consequence, the account released 10 albums between September and December 2025 and landed three songs on Spotify's Viral 50 USA playlist. One of those tracks even ended up soundtracking a Selena Gomez Instagram post. Then Deezer publicly stated that her music was completely AI-generated.

Maybe the vocals are a little too smooth. But the music? Damn good. And that's what makes it scary. Thousands of people loved those songs without a second thought, and honestly, I get it.

This isn't a niche problem either. In July 2026, Deezer reported that AI-generated tracks made up more than half of all new music uploaded to its platform, around 90,000 tracks a day at its peak. And a Deezer survey found that 97% of participants couldn't tell the difference between fully AI-generated music and human-made music.

So if you can't judge the song, what do you judge?

The artist behind it.

When I found that artist on my YouTube Music radio, I didn't need a detector. I just looked at everything around the music, assumed it was AI, and moved on.

At the end of the day, we must take initiative and vet them on our own.

My 4-Step Process for Vetting an Artist

There's an order I follow every time an artist makes me suspicious.

Each step builds on the last, and by the end, I usually have my answer.

Step 1: Read the Bio

I always start here. A real artist's bio gives you some perspective on their career. Where they're from, what shaped them, who they've worked with, what they're chasing.

If the bio reads like they came from outer space, or from some dreamy, made-up place, and tells you nothing real about their journey, I immediately start questioning them. Vague poetry about "analog textures" and "healing" with zero names, places, or milestones is a red flag.

You know when something is bullshit. The bio is the first indicator.

Step 2: Check the Discography

Next, I look at how much music they've released and how fast. Humans need time. We have to live life before we can write about it. Most artists take at least a year or two between albums.

If an artist is busting out album after album with no break, it's clear to me. It must be AI.

Remember, Sienna Rose put out 10 albums in about three months.

No human is doing that.

Step 3: Study the Visual Identity

This one is my favorite, because it comes down to something beautifully human: we're inconsistent.

We change. Our taste evolves. Our art evolves. My own visual identity for TLR Sessions has changed drastically from Session No. 001 to Session No. 012. That's what happens when a real person is growing in real time.

Now, some artists do keep a consistent look on purpose. Think Jamie xx, who has two albums with the same style. That's an intentional creative choice, and it's rare.

What I question is:

  • Artwork that looks like AI slop (You already know, just get on instagram and look at local restaurants using AI)

  • Every single release looking nearly identical, like it came off the same assembly line

  • Press photos that don't look like a person you could actually bump into at a show

Step 4: Google Them

Then I Google the heck out of them. I'm looking for interviews, live performances, and coverage from vetted sources that have reviewed their work. Real artists leave a trail: a local show flyer, a podcast appearance, a blog interview, a friend tagging them in a photo.

If I can't find a single trace of a human outside the streaming profile, that pretty much answers all my remaining questions.

Putting It Together: Loom Room

Loom Room is a great example of an artist that checks every box on my list. The bio, the release pace, the visuals, and the lack of any real-world trail all raised flags for me. A crowdsourced AI artist list called SpotAI has it marked as "validated as AI." Some listeners argue there's a real producer behind it who used AI for the album art.

And that's kind of the point.

When you have to dig this hard to find a human, the burden of proof isn't on me anymore. It doesn't pass my vetting, so it doesn't make a Session.

Where I Draw the Line on AI in Music

I'm not going to act all high and mighty when it comes to AI. I use it myself. It helps me understand things that are unknown to me and figure out what I need to do to get something done. But it doesn't think for me, and it doesn't execute my ideas.

That's how I see it for artists too. If a musician uses AI as a tool in their process, to help with a mix, brainstorm an idea, or learn something new, that's fine with me. There's still a human with a vision steering the whole thing.

What I think is outright crazy is fully generating a song, putting it under a made-up artist, blowing it up, and making money from it. It feels soulless.

But I'm also realistic. We're not going to be able to run away from this. It's our reality now. So the question isn't whether AI music exists. It's what we choose to support.

Why It Matters: Music Is an Entryway Into Someone's Soul

I want to support human beings who poured their heart, emotion, time, and energy into a project. That's the whole reason I do this.

Music is an entryway into someone's soul. When a musician creates something so beautiful and magical that it stops you in your tracks, part of the magic is knowing a person felt that first. When you find out no human made it, it loses its spark. Its spizazz. The thing that makes us curious to dig deeper, read the liner notes, and find out what they were going through when they wrote it.

I pursued acting for a long time, so I think about it like this. If I found out Viola Davis was AI, I wouldn't be able to connect to her story or her emotions anymore. It would take the curiousity away. The performance only hits because a real person lived something and brought it to the screen.

There's a name for that weird feeling: the uncanny valley. It's what happens when something looks almost human but slightly off, like the characters in The Polar Express. Your brain can't quite trust it. That's exactly how it feels when I find out a song or an artist I was vibing with is AI.

Everything I felt suddenly has nowhere to land.

What Curation Means to Me Now

For me, curation comes down to trust. When you open a TLR Session, you should know that I'm sharing music made by humans, with a whole feeling attached to their work.

I don't want to share a song that doesn't have a story to go with it. So I vet every single song that ends up in a Session, every artist I write a portrait about, and every person I invite to be apart of The Listening Room

That's my promise to you, Roomies. My music is vetted.

And if something ever is AI, you'll know, because I'll tell you.

What to Do When You Get That "Wait, Is This Real?" Feeling

I want you to have the confidence to sniff out AI music yourself. And more than that, I want you to feel confident finding artists you actually want to support, knowing there's a real human on the other side who took the time to pour their heart into that album.

So the next time an algorithmic playlist hands you a song you love and something feels a little off, run through the four steps: bio, discography, visual identity, Google.

If It's AI

My advice is not to share it and not to keep streaming it. Every stream is a vote for what gets made next. But hey, to each their own.

If It's a Human You've Fallen For

This is where the fun starts. Show up for them:

  • Buy the merch

  • Go to the show

  • Comment on their posts and engage with what they share

  • Introduce them to a friend who hasn't heard them yet

That last one is my favorite. It's what we call The Afters here at TLR: what you do after a song moves you. Real artists grow because real people pass their music along.

Listen Like You Mean It

AI music isn't going anywhere, and it's only going to sound better. But you don't need a perfect ear to protect what you love. You just need to be a little curious about who's behind the music.

That's what I do for you every week. If you want songs from real humans with real stories, vetted before they ever reach your ears, Sign the Slip and join the Roomies.

A new TLR Session lands in your inbox every Thursday.

Sources

Join TLR Sessions

Contact

Info@TLRsessions.com

661-863-7423

© 2025. All rights reserved.