I have never been good at describing my music taste to other people. If someone asks, I say something vague like “a bit of everything,” and we both quietly agree to change the subject. I do not know the names of most genres I like. I just know what sounds right at a given hour of the day. There is something dancier for the late afternoon, something quieter for when the light goes flat in the evening, something else again for an early train. I could not tell you what genre most of it is. I could tell you when I want it.
So the other evening I copied my liked songs on YouTube Music, handed it to ChatGPT, and asked it to tell me what they say about me. I had been curious for a while whether the shape of a likes list could really say anything true about a person, or whether it was the kind of question an AI would only ever flatter you on. I expected something generic. I got something more pointed than that.
A rough sample of what I gave it: James Blake. Khruangbin. Fatboy Slim. Bon Iver. James Brown. Moby. Olafur Arnalds. Maribou State. NEU!. Nina Simone. Hot Chip. A handful of long live sets. Some ambient pieces. No tidy pattern I could put words to. If you had asked me to draw the shape of my own taste, I could not have done it.
Here is what I typed:
“Above are my liked songs on YouTube Music. What do these say about my personality? Be honest. Don’t flatter me. Make your answer detailed.”
I hit send. There is a small embarrassment in handing your music to a machine and asking what it sees in you. You are aware, even before the reply comes back, that the question is partly vanity and partly real.
The reply was longer than I expected, broken into sections like a school report. Most of it was readable in the way ChatGPT can be readable — competent, balanced, sometimes too polished. But four passages really stood out.
The first was pretty revealing. “You don’t seem like someone who listens mainly for catchy hooks. You seem like someone who listens for states of mind.”
That felt true in a way I would not have written about myself. I do choose songs by mood more than by melody. I rarely care if a track is on the radio. I care if it fits whatever weather is going on in my head at that moment. I sometimes reach for a song the way a person reaches for a cardigan — because the temperature of the day calls for it.
The second observation was structural. The list, ChatGPT said, has “two poles”. One pole is the still, late-night, reflective stuff — Bon Iver, Olafur Arnalds, Sleeping At Last. The other pole is groove and motion — James Brown, Jungle, Fatboy Slim, Maribou State. The read: I want both. I want stillness to think in, and rhythm to get out of my head with. I had never put those two halves of my listening into the same sentence. Once they were in the same sentence, it was hard to argue with.
The third line was tougher, and pointed at something I had not named for myself. “Sometimes melancholy can become a kind of identity furniture. Familiar. Aesthetic. Hard to leave.”
I am going to be thinking about “identity furniture” for a while. The implication I suppose is not that I am sad. The implication is that I might be a bit too comfortable with sadness as an aesthetic — that a certain kind of pretty heaviness has become part of how I decorate the room. That is a different and slightly more uncomfortable observation.
The fourth was the bluntest. “But the danger is that you may spend too much time curating the emotional atmosphere of your life and not enough time doing the plain, repetitive, unsexy things that would actually change it.”
I did not enjoy that one. It arrived politely and sat in the room for a while. It is not really a diagnosis. It is more of a nudge — the sort a friend who knows you a bit too well might make, in a low voice.
Read this before you try the same thing
ChatGPT is not a psychologist (nor am I) or a music critic, and it is not qualified to diagnose anything about your personality. This post is for entertainment. If you decide to run the same prompt against your own likes list, treat it as a lighthearted, curiosity-led exercise. Don’t read it as a diagnosis, and don’t make any actual life decisions based on what an AI says about your playlist.
What an exercise like this can do, at best, is hold up a mirror. Whether the reflection looks anything like you is a separate question. Some of what ChatGPT said about me was generic enough to apply to half the people I know. Some of it I disagreed with as soon as I read it. A few sentences did land, and they landed because they named something I had not quite named for myself.
That is the most a tool like this can do. It does not know you. It knows the inputs you gave it. The work of deciding what is true is still yours.
I went back to my likes afterward. The first song I played was “Feel Good” by Khruangbin. I do not know what that says about me. I do know I liked it.
How Sharp Is Your Era Memory?
Every memorization style can reflect a different way of holding the past—through feelings, stories, details, or senses. This beautiful visual quiz reveals how your mind naturally stores what matters and what that says about the way you experience life.
✨ 10 questions. Instant results. Guided by shaman Rudá Iandê’s teachings.
