AI elevator music

Fake lo-fi beats to relax/study to.

21 Apr

Terry Nguyen, Dirt's senior staff writer, on AI-generated "elevator" music and media layoffs. Plus, some weekend podcast recs ft. the girl-mosses of Dirt.

If you are a Spotify listener who gravitates to a genre of “ambient” music, say, meditative dubstep for deep focus, chances are, you’ve probably encountered some strange, recurring tracks. These “songs” are unimpressive and short, roughly one minute long, and produced by a small, unknown artist with a few tracks to their name. These aren’t real songs. They’re generative tracks, “mood music” produced by AI generators and uploaded by Spotify. Not only that, Spotify seems to be strategically inserting these generative tracks into popular ambient/chill/instrumental playlists to rack up streams.

Earlier this week, the producer Adam Faze noticed an instrumental track that kept recurring on his Spotify radio. The same song would resurface under a different title and artist. Faze compiled all 49 examples into a playlist titled “these are all the same song,” but there are many more examples of this filler music floating around in the Spotify ether. When I listened to the playlist, a phrase came to mind: elevator music. The track that caught Faze’s attention isn’t the smooth brain jazz that we typically associate with elevator music; it’s a classical-ish piece that had a theme-y feel to it, like it could’ve soundtracked a movie or a video game. One version vaguely reminded me of the Pirates of the Caribbean theme song.

Spotify is only inserting this kind of music into playlists where they’re “more likely to get away with it,” Faze told me over text, because people have been “passively listening for hours on end.” The tracks are short, too, for a reason: By the time you realize it’s a song you’ve heard before and move to skip it, it’s almost over.

Why is Spotify doing this? For revenue—because it has yet to turn an annual profit. Now that podcasts are no longer a surefire money-making avenue, Spotify is pivoting, or rather, leaning into a strategy it’s kept in its back pocket. The streaming platform has surreptitiously hired producers to compose filler music since 2016. These tracks are then licensed to the platform at a lower rate than tracks from major labels. The rise of AI music generators has only made it easier for synthetic music—and media, more broadly—to proliferate.

Musician and music writer Jaime Brook has a bleak (but good) prediction on the effects this will have on the music/streaming industry:

“The unavoidable outcome of all this is the further devaluation of recorded music. The major labels and the streaming services may have conflicting goals, but they are collaborators, not enemies. They have already conspired to devalue recorded music once before, when the labels embraced the all-access subscription model as a means of protecting their intellectual property from the existential threat of digital piracy. The threat posed by generators will inspire a similar gambit.”

IN OTHER (BLEAK) NEWS

  • On Thursday, Insider announced it will lay off about 10 percent of employees across various departments. The layoffs are reportedly unrelated to the newsroom’s adoption of ChatGPT, which Insider EIC Nicholas Carlson encouraged in a memo to staff last week.

  • CEO Jonah Peretti effectively shut down Buzzfeed’s news division, laying off around 180 employees. Peretti offers a half-apology for the disastrous SPAC-ing in 2021, says he has “learned from these mistakes,” and says there will be a “number of select roles” open at HuffPost and BuzzFeed Dot Com for BFN employees.

  • Buzzfeed News has been home to some truly phenomenal pieces of journalism. Aside from its big Pulitzer Prize-winning (and nominated) investigations, I’ve personally enjoyed works from BFN’s tech/culture desk, from reporters like Katie Notopoulos, Scaachi Koul, Elamin Abdelmahmoud, Steffi Cao, and Kelsey Weekman. Solidarity with all the media workers who’ve been affected!

Use this to separate content

PLAYLIST

Reports of the podcast's death have been greatly exaggerated...

  • Daisy joins Matt Rodbard on The TASTE Cooking podcast to talk Web3’s impact on the restaurant industry, celebrity mukbangs, and being “post-food.”

  • CNN’s Audie Cornish had me on The Assignment with TikTok trend forecaster Agustina Panzoni​​ to talk clickbait fashion, MSCHF’s Big Red Boots (which Emily Jensen wrote about for Dirt), and the state of trends.

  • I joined Slate’s Rachelle Hampton on ICYMI to talk about how plastic surgery has shaped beauty trends, Blac Chyna’s de-transformation, and more.

Use this to separate content

MIXTAPE

Good links from the Dirtyverse.

  • What was Twitter, anyway? A [redacted] number of think-pieces later, I think we won’t know so long as the site is still around and active. (NYT Magazine)

  • Burning (2018) director Lee Chang-dong is also a short story writer. He has a piece out in the New Yorker, which I enjoyed.

  • Succession’s Willa Ferreyra has had a First Lady-worthy transformation. (Coveteur)

  • Reddit wants to get paid for the data it provides AI models. (NYT)

 🌱 JOIN THE DIRTYVERSE

  • Join our Discord and talk Dirt-y with us. It’s free to join! Paid subscribers have access to all channels.

  • Follow @dirtyverse on Twitter for the latest news and Spotify for monthly curated playlists.

  • Shop for some in-demand Dirt merch. 🍄