Exploring AI and Superfan

“카톡!” (Ka-talk!)

My phone rang in my pocket. It was a text from my older brother, Collin, on KakaoTalk, a popular Korean chatting app.

“Get this app,” he messaged, followed by a link to an app called Superfan.

After doing some cursory internet researchI found that it was an app that used an artificial intelligence to analyze your listening history on Spotify and, using the information it gathered, conjured an oddly specific description of your apparent vibe for the week. Using Superfan, you can also share these descriptions with your friends by creating a group.

Opening the app and asking for that week’s recap, I was told that my vibe was “like a summer day spent dancing in a field of wildflowers.”

“Huh,” I shrugged to myself, “that’s interesting.” Scrolling through the statistics the AI used, such as total listening time throughout the week, top songs, top artists and albums, I gradually realized that maybe, just maybe, this AI was onto something.

Wasia Project, Faye Webster and Taylor Swift were my top artists that week. To test the AI’s knack for producing such an oddly specific vibe, I decided to relisten to my top tracks of the week: “Petals on the Moon,” “Right Side of My Neck” and “invisible string.”

Closing my eyes, I could imagine and almost feel the warmth of jubilant sunbeams and the pointed stems of wildflowers grazing against my skin.

I was intrigued and probably under the spell of the priming effect. But I had to let everyone know.

Furiously texting my friends about Superfan, claiming Collin’s discovery as my own in the process, I told them all to download the app and then once they did, asked them about their vibes of the week.

“What the fuck!? Mine says ‘crying over your ex while simultaneously trying to move on from them.’”

Oh Superfan, you didn’t have to expose him like that.

Superfan has become a fun way to stalk what music my friends are listening to (and what they could be going through in life) — it has become a new way of passing the aux, without the aux. Normally, I would be afraid of being the one with aux privileges: Who knows if everyone else will vibe with my music — not everyone feels like dancing in a field of field flowers. But Superfan strips aux anxiety since anyone can lurk on what you are listening to and see what your oddly specific vibe is. 

Nevertheless, I’m still skeptical of AI, especially when talking about the arts. You are essentially combining a cold, calculating computer of some kind with something so emotional and complex and human —  how we individually listen to music. Music is often an outlet for people to find solace and comfort with their emotions. AI’s ability to mimic human creation and creativity blurs the lines between human and machine art, possibly to the point where it’s impossible to tell what is AI and what is not. Considering that AI refers to works created by real people to generate their musical arrangements, controversies are brewing.

Understandably, Superfan using AI to merely generate prompts versus some of these artist AI’s generating full original songs is a large leap to make. Spotify similarly has an AI-generated playlist called “daylist” — your day playlist. The AI will generate different titles every morning, afternoon and evening based on your moods from your listening patterns. But most of the time, I can hardly decipher these titles. For example, yesterday morning, my daylist was titled “yearning downtown vibes morning,” and today it was titled “microcore early Tuesday morning.” Worse, my brother’s daylist was “rizz gyat evening” — what songs could possibly fit that vibe?

Maybe I haven’t listened to enough music to truly enjoy and understand these niches, but the AI has probably been fed trendy buzzwords to entice their users with vague niches. Since I, and likely many others, cannot understand some of these titles, I assume everyone gladly accepts these daylists since they’re filled with familiar music anyway.

Superfan and Spotify’s daylists add an innocent, nice little social aspect to everyone’s musical ice cream flavors with AI sprinkles and cherries on top. Add anything more, your musical sundae might just melt and topple. But recently, AI generative software might have just started to bring out their flamethrowers.

AI has improved dramatically in the music industry. With plenty of vast datasets from different genres, controversies emerged from AI potentially breaching copyright issues, and more importantly, stripping artists’ voices and creativity without compensation. Just imagine, you’re a musician who spends years perfecting your craft. You write songs, pour your heart into the lyrics, and spend hours in the studio getting everything just the Goldilocks amount of right — you’re Grammy-bound. Now, an AI that can analyze thousands of songs, learn the patterns and structures and then generate new music in minutes that sounds eerily similar to your style or even the styles of famous musicians swoops in, outshining your work — it’s the Terminator. AI has even gained the ability to revive artists from the dead, able to puppeteer their voices to sing covers of modern songs posthumously. As much as AI is amazing — almost magical — it is important to not overstep what has defined our art since the very beginning — humanity.

So why not keep it simple and innocent like Superfan?

For a while, anything created by AI was deemed fake: See a painting created by AI? Oh that’s not real, that was made by AI. It’s fake. Listen to an AI cover of Frank Sinatra singing Smells Like Teen Spirit? That’s scarily beautiful, but I know Sinatra died a while ago so again, that must be fake. But now when I listen to a Drake and The Weeknd AI collab, I think it’s real. Without any preface that any song was made by AI, it might be impossible to tell whether a current artist actually produced and sang a song.

AI is terrifyingly beautiful — it is filled with so much potential but equally with potent destruction. So I’ll stick with my little playlists dancing in wildflowers or pretending to be a coffeehouse musician while actually binging reality TV (please add me on Superfan).

Statement Columnist Philip Ham can be reached at philham@umich.edu.

The post Exploring AI and Superfan appeared first on The Michigan Daily.


Comments

Leave a Reply

Your email address will not be published. Required fields are marked *