when it walks like a duck, and quacks like a duck, yet is absolutely not a duck
IN HER UNIVERSAL HEARTBREAK ANTHEM, “White Flag”, Dido takes us somewhere we’ve all been, but never wanted to go. Her soft voice belies the real weight of a song with eternal resonance. Hers is a rousing refusal to leave limp love. She sings in full surrender to the chance of love persisting, even if that proves delusional.
When the song rips through speakers, it broods, it stings, it spreads its eerie clouds over the airwaves. She doesn’t know if the door is fully closed, but if it’s ajar and love’s not too far, she will be there. Her love won’t end for wont of trying.
When she sings:
“I will go down with this ship
And I won’t put my hands up and surrender
There will be no white flag about my door
I’m in love and always will be”
We feel the intensity of her defiance. We remember when we too felt the sting of warm love falling away from our fingertips. We are transported to our teenage selves feeling heart-shaking, nervous-wrecking, adolescent infatuations.
That’s a ballad doing what it’s supposed to do: ripping us out of our comfy seats, softening our hearts, and giving work to idle tear ducts. As we place our ourselves in the shoes of the songwriter, we find words to describe our own melancholy or joy or rich triumph.
Now, what if you found out your favorite ballad wasn’t written by a person? What if “Someone You Loved” by Lewis Capaldi were written by ChatGPT? How would you feel? Would it change your appreciation of the song?
Or is it all about how it sounds? A banger is a banger regardless of who, or what wrote it?
Ballads rely on emotional intensity and vocals, so let’s imagine another kind of song. What if it’s something more vibey? A song on your summer playlist that instantly makes you smile. Suppose you discover that while the vocals are human, most of the production (writing, instrumentals, sound engineering, mixing, etc) is AI-generated. How would you feel listening to it?
Imagine it’s your last night on vacation in Mexico City. You’re drinking mezcalitas with your friends at a bar with orange lights and rust-brown outdoor seating. The DJ plays this song that rips a roar from the crowd. The ground begins to swell as people lend their bodies to the dance floor. You haven’t felt this need to boogie since the first time you heard Jungle. You rise to dance, and your friend rushes after you, but before the song is lost forever, she Shazams it. “JAKE BANKS”, she clicks on this Spotify profile wondering what else he’s dropped only to find an “AI vocalist” banner. Would she feel bummed out?
II.
I READ WITH HORROR the news that an AI-generated song topped the Billboard Country US charts last month. My horror morphed into embarrassment when I found myself bopping my head to “Livin’ on Borrowed Time“, a song by Breaking Rust, an “AI band”. I later found out the song has been streamed more than eight million times, and I’m responsible for two of them. I’m not a country buff, but to my untrained ear, the song sounded good.
I have so many questions. Who is making this stuff? Why invite AI into the music room? How is AI used in music production? Does it work like AutoTune or synth beats, extending the capabilities of musicians? Or is it a plastic (and potentially illegal) attempt at mimicry? My intuition tells me this is Soylent trying to replace whole foods, should I trust it?
Who is better off because AI music exists? I doubt it’s listeners. Music is not scarce; there is far more good music available today than I will ever have time to listen to, and finding a new artist is a rare joy in life. Part of the fun is digging into their discography and hearing how their sounds and influences change over time. If the artist isn’t human, then is there a story? It’s just statistics and cardboard.
Many of the pivotal moments of my life are punctuated with lyrics from songs I have on repeat. “What’s the point of flying first class on a crashing plane?”, echoed in my brain for months before I went on sabbatical. When I surrendered and accepted that I was a writer, and started working with the inimitable Rachel Jepsen two years ago, “It’s gonna take a little work….now that you’re here, work!” by Charlotte Day Wilson kept my head above water. When I got mild Substack fame here last year, Cleo Sol’s lyrics featured prominently in the story.
If I found out any of those were written by a machine, I would feel let down. And I know I’m not the only one who feels this. My heart tells me AI music is unethical, but then I heard a song on Suno that I liked. I did not want to like it, but I did. Realizing this felt bizarre—it was like eating the smashburger from Lovely’s and trying to convince my brain it wasn’t tasty.
If that song played on my Spotify shuffle, I would have no clue it was made with AI. Might AI music become one of those things we publicly say we don’t like, but we actually don’t mind consuming, like fast fashion or plastic water bottles that are very unpopular in social settings, but if convenient enough, we consume when nobody’s watching.
But music works differently than utilitarian goods. The bigger question is:
Why invite AI into music at all? Can’t we have any sacred rooms in this house where we cherish human creativity for what it is?
Making music today already seems accessible. There’s tons of YouTube tutorials for every instrument. Public libraries in the US host free guitar and bass sessions, and offer software and books for people of all levels to learn music skills. Online music schools like ArtistWorks let you learn from world-class musicians.
AI is supposed to kill the mundane stuff from our lives: read our leases and contracts, and find convenient times for our meetings—so we can spend our time making music, cooking full-day braised beef, and sending physical postcards to our people.
It should not steal good things from us. In creative work, the process is as important as the artifact. Making music, or writing, or mastering photography is often challenging, but creative struggle and mastery are good for us, and we should encourage more, not less of them. I haven’t tried the Suno platform myself, and I don’t know how much creative wiggle room it gives to creators, but I’m wary of the CEO’s comments that “people find making music unenjoyable because of the effort required.” My impression is that musicians find meaning in the practice itself, and they want the satisfaction of developing skill over time. And it’s ok if everyone isn’t a musician.
Writing this essay wasn’t easy. I wrote forty or fifty rambling openers for this piece, each time inching closer and pulling farther away from the message I wanted to convey. I could’ve asked Claude to spit out an acceptable opening. But if I did, I would have missed the moment when everything clicked in my head, when I finally figured out what I was trying to say and how to say it. That’s something I deeply treasure. I imagine making music works in a similar way.
What then should we do?
Prioritize human music. There’s an entire universe of incredibly talented musicians out there, in your city, hanging in your Spotify app, maybe even in the jazz bar a few blocks from your bus stop, waiting to be discovered.
I do think AI music will eventually get so good it permeates through social media, through our favorite cafes, bars, and restaurants. But some people will reject it. IHeartRadio just banned AI music and synthetic DJ’s from their station as part of their “Guaranteed Human” promise. I’m not against AI music existing. I just think there should be reasonable disclosure, and fair payment to musicians whose work was allegedly used to train those models.
Independent musicians already struggle to make real streaming revenue competing with other humans, and AI Vocalists and their farms of listening bots are unlikely to sway things back in their favor. There are real lawsuits and legal disputes being hashed out in the courts around copyrights, fair use, and licensing deals. Things are going to get a little weird until things get sorted.
Music means many things to people. It’s your heartbreak anthem when you need to feel all the feels. It keeps your heart pumping at 140bpm during your workout. It transports you away from the guy elbow jabbing you on the bus during your commute. Listening to the playlist your boyfriend made you can reconnect you even over thousands of miles of distance.
AI Music might walk like a duck, quack like one, even waddle in webbed feet, and that might suffice or even exceed our expectations in many cases. But if you want that soul-deep, butter-rich, human connection, and want to appreciate decades of craft, I think you want the real deal. I think you want the one you can see on tour.
This is my last AI post for the year. I’m working on something much more exciting coming soon. It’s more similar to kind of writing I did last year, and I am excited for you all to read it :)
I went through this too, only I knew it was AI music going in and I still had to admit how much I liked it (‘90s/‘00s grunge and rock hits composed in the style of old soul/RnB by Professor Nick Harrison—a real guy but created with AI right down to the voice I think.) I had this playlist on repeat one weekend, and I noticed after the sixth-or-so listen, it didn’t hit as vibrantly. It sounded more hollow. The novelty wore off and there was no “realness” to support it. I’m hoping we’re all just in this ecstatic state of novelty where we want to try ALL the things, but it’ll eventually wear out, regulate and best-practice itself, much like any new technology. We created nuclear bombs and somehow we’re still here???
Looking forward to your new stuff next year!
It’s interesting, I feel a big differentiator is that with human made music, you know there is extra depth hidden behind each lyric, and each layer of the beat. And can only hope to interpret as the beholder, with your relationship to the words changing over time.
I’d assume ai music lacks this depth because there is less intentionality, but maybe that’s just something I tell myself for comfort. After all, there are things it notices in the minutiae that humans don’t recognize at first glance. Maybe someday it too will give depth that has artistic meaning not differentiable from any other musician or painter