My affair with Spotify’s AI DJ

Spotify’s AI DJ could be a glimpse at the future of consumer-based AI products.

I’ve been having an affair with an AI.  

It started last year at my brother’s wedding in February 2023. We were sitting in F. Scott Fitzgerald’s childhood living room in St. Paul Minnesota. You can book it on Airbnb now apparently.  

U.S. based company OpenAI had just launched ChatGPT the previous November while DALL-E’s text-to-image art renders were sparking discussions on art, morality and originality within AI. Creatives around the world were concerned about their livelihoods and artistic integrity.   

An article had just come out in the New York Times about a hallucinating new chatbot that declared its love for columnist Kevin Roose and requested him to divorce his wife. My dad was laughing into his phone screen, in the dim Fitzgerald home lighting, reading this article to us. 

What’s the point of it all? We wondered back then. It seemed exploitative. AI was this nonstop wave of news, media and discussion. My brother saw it as an unnecessary evil. As far as I was concerned, I felt content existing without AI. I told them I wouldn’t have something like that in my life. How creepy it is. We talked about it as a group in the living room before we had to pick up the wedding cake at the bakery.  

That same month, Spotify released their obligatory implementation of AI with its new AI DJ service. Spotify acquired AI voice generation service Sonantic to create “DJ X” using the real voice of Spotify’s Head of Cultural Partnerships, Xavier “X” Jernigan.  

I tried DJ X that same month, and it was just supposed to be a one-time fling. I used the feature more and more as the months went by after the wedding.  

Instead of a playlist to choose from or an album to pick, it was enticing to just have an artificial-intelligence-deep-language-learning algorithm with a very convincing voice, powered by Spotify and their intensive data collection, pick all your songs for you.  

DJ X knows what you listen to on Thursdays. He tries to guess the emotions that certain songs might evoke from you. He notices your tendencies to listen to rock on Friday nights. Sometimes he’ll introduce an entire genre, ranging from the specific to the mundane. He might play only Detroit Punk bands for an hour or only music from artists chosen by staff curators the next.  

I felt a little dirty every time I used DJ X, like I was using something I wasn’t supposed to. I started to use DJ X every time I got in the car. I got mad at him when he wasn’t playing the type of songs I wanted to hear. What does he know about taste anyway?  

There’s a blue button to press if you don’t like the vibe DJ X is going for. Press it and he’ll change his song direction or decision-making process. I’d hit that button over and over and over, just to see what he would do.  

Spotify’s AI DJ X in action, playing “Lego Ring” by Faye Webster and Lil Yachty.

Classical. Folk Acoustic. New Wave. Heavy Metal. Industrial Noise. House. I liked more than I didn’t. I kept pressing that button.  

Through awkward pauses and mispronounced band names I found something in DJ X. He was an AI amalgamation of data propped up by human art created by screaming voices in every font and color. I wanted to objectivate DJ X, culminate this experience in this software concretely. I wanted it to feel like a robot. Something about it all felt creepily unhuman. Like a robot wearing human skin. 

There’s too much context to robotize the consumption of music. These rhythms feel intrinsically human, electronic instruments or otherwise. Through every vibe I skipped I found something I knew or someone I loved. Why is it playing Smashing Pumpkins? That’s what my dad listens to. Another K-Pop song? That’s more my girlfriend’s taste. I haven’t listened to that song since high school, why would I want to listen to it now? And it controls your heartbeat regardless.  

 What’s the point of it all? I still wonder. Is it just consumption fuel for never-ending revenue? AI has entered many art spaces in the past year in ways big and small. It is growing and it is learning. What might this integration look like ten years from now when this technology advances? When the DJs actually play what you wanted to hear, because they know you better than you do.