Over the last couple of months, in protest of CEO Daniel Ek hurling venture capital at the AI weapons company Helsing, a wave of musicians have stripped their music off Spotify. Indie faves like Xiu Xiu, Deerhoof, Hotline TNT, Godspeed You! Black Emperor, and King Gizzard & the Lizard Wizard have all joined the fray. Last week, Deradoorian and Massive Attack announced their departure from the platform, quickly followed by Sylvan Esso. Then, out of nowhere, on Tuesday, Ek announced his departure, stepping back but not exactly stepping out. He’s moving from CEO to Executive Chairman, and the company’s former CTO and CBO will become the new co-Chiefs, reporting to Ek.
It feels like the latest and grandest move in the streamer’s ongoing reputation rehabilitation campaign. During the last few months, the company has unleashed a flood of features and security measures that might just be conveniently timed, but also feel like tactics to deflect from the negative PR. Among them is a tool to let users DM friends on the platform—a feature the company reheated from a decade ago, when they tested then killed it after barely anyone used it. There’s also the introduction of lossless audio, which will let users listen to higher-quality versions of music.
The only update that actually hooked me is Spotify’s new announcement around AI and spam. They claim that over the last year, they’ve removed 75 million songs engaging in what they call “spam tactics”: people mass-uploading generic nonsense; the same songs uploaded twice or thrice; cheats to hijack the SEO (like stuffing keywords in titles to get algorithmically surfaced easier); and “artificially short track abuse,” which is when people split up longer songs into short segments to rack up royalties. The company teased a new “spam filter” that will help Spotify avoid recommending songs engaging in these slimy tricks going forward. They also announced a new tool that artists uploading music to Spotify can use to disclose if they’ve used AI.
Clearly this about-face is about saving face. Spotify’s emphasis on playlists and endless streaming has prioritized passive-listening background music in a way that’s directly facilitated spam and AI. As writer Liz Pelly discovered, the platform partnered with companies who hire songwriters to pump out stock tunes basically indistinguishable from AI, letting them release their music under a slew of fake “ghost” profiles. Spotify would cut deals to pay them less royalties, then stuff these conveyor-belt tunes on curated playlists, thus maximizing Spotify’s profits. Earlier this year, Spotify announced a partnership with voice generator ElevenLabs for AI audiobooks. We can’t forget about Spotify’s heavily hyped AI DJ, an algorithmic tour guide with a “stunningly realistic voice” that many people say barely works (on a recent car trip, a friend prompted it to play ocean-themed music and it played Frank Ocean). Ek himself has glazed AI, saying in 2023 that it’s “great culturally but also benefits Spotify because the more creators we have on our service, the better it is and the more opportunity we have to grow engagement and grow revenue.” In Spotify’s contextless vortex of passive consumption, pleasantly inoffensive AI slots right into the algorithm, infinite fodder to keep users chained to the platform.
It’s a half-baked half-measure. Similar to YouTube’s self-disclosure system, which lets creators mark if their video involves AI with an “Altered and synthetic media” label on the video, Spotify doesn’t seem to be making this a mandatory admission. Spotify’s clearly trying to frame the update as empowering for AI merchants, a way for “responsible” musicians to “share if and how they’re using AI,” as if it’s a point of pride. As Spotify’s Global Head of Music Product Charlie Hellman told Variety, they’re “not here to punish artists for using AI authentically and responsibly. We hope that they will enable them to be more creative than ever.”
Over the last year, Spotify has become congested and infested with fake AI profiles: accounts with fictional names and airbrushed avatars posing as real artists while churning out robo-slop. There was the Velvet Sundown, a psych-rock “band” that amassed over a million monthly listeners before the creator admitted it was something like a “social experiment”—no band ever existed, and the photo was as fake as the music. There was the case of Charleston rocker Johnny Delaware, a real human musician, who was vexed after the podcast Switched on Pop and an Atlantic writer questioned if he was AI thanks to Spotify’s poorly programmed custom playlists. Add to that all the lo-fi beatmakers who told me they’ve lost thousands of dollars after being replaced by what they believed were AI musicians on the platform’s official playlists. While other platforms like Apple and YouTube Music are also swamped in AI waste, Spotify is in some cases actively promoting it. See the “epic fantasy” of droid-rocker Echo Harper, whose avatar looks like Call of Duty fanart and who has a song in the top half of Spotify’s curated “Just Rock!” playlist.
Many of the platform’s current biggest AI blights wouldn’t, by the standards of Spotify’s new announcement, be considered robotic spam or subject to removal. There’s an emerging middle-class of AI acts with hundreds of thousands of monthly listeners that have official “This Is” playlists, which Spotify uses to compile the artists’ biggest songs. Spotify has done it for David Sven, an artist whose bio promises “soul-rock for people who’ve felt too much.” Propping up these anti-artists promotes a kind of false idolatry; undiscerning listeners read their banal backstories and watch their flashy videos and soon become active devotees. We’re living in an era of mass misanthropy—an antisocial insanity ushered in by COVID and screen ubiquity that’s led people to treat ChatGPT like a lover, therapist, and bestie rolled into one. The music just adds another layer of unreality. Prepare for QAnon levels of musical hysteria, where your religious grandparent says their favorite musician is the pseudo-psalms act Holy Groove.
Let Babylon Burn, a reggae “artist” sculpted from the ether, recently cracked the top 10 on Spotify’s Viral 50 charts for Norway, Germany, and Switzerland with the uncanny outpouring of “I Forgive That Man.” The music has tricked a horde of people who write YouTube comments like “Can’t stop listening, this should be an international hit” and “I am 44, and with this song I pass through all the most tender, intimate moments of my life.” You can’t really blame them—besides the nonsensical music video where the singer doesn’t sync up with the vocals at all, the only overt tell that this is AI music is the Spotify bio, which is buried at the bottom of the page and describes the style as “heartfelt, self-written lyrics with a unique fusion of acoustic music, production, recording, and AI-enhanced creativity.” So in effect, Spotify’s AI self-disclosure will do nothing: These people are already churning out vacuous robo-reggae and admitting it.
The bio of another AI artist, Enlly Blue, begins with the exact same line, before going on to flex “recordings enhanced with cutting-edge AI technology,” so “every track becomes a bridge between heart, sound, and imagination.” Spotify has a “This Is” playlist for Blue, whose cover photo looks like a fairy-dusted Marilyn Monroe. Blue just crossed one million monthly listeners, an astonishing amount for some lines of code. Their vacant dreck “Through My Soul” is all over official Spotify Viral 50 charts across Europe, along with ZUNRAID’s “Whiskey morning,” which has a Suno demo sitting online. The irony of soulless soul would be funny if it wasn’t such an egregiously overpolluted un-genre; there’s an abyss of these fake artists begging for deliverance through nasty purple prose and botched metaphors about ghosts and secrets. Have a glimpse at Royal Blue Notes, a page with almost 200,000 monthly listeners that’s dropped seven EPs in the last couple of months; they deploy AI to “bring the spirit of B.B. King back to life!”
Spotify won’t prohibit this music—not because it thinks it’s innovative or ushering in a new era of technological futurism (the platform has never cared about culture) but simply because it’s generating streams. If the company actually wants to be transparent and control spam, they need to go way harder. Clean up the whiskey blues botfarms, the city slop, the “romantasy” by rote fu-metal, the machine-mangled MAGA porn of Full Metal Patriot, who offers “no labels, no compromises, just patriotism cranked to 11” and music made with “AI + Soul.” Remove these grotesque counterfeits, or at least adorn them with Scarlet Letters of Shame—I’d like to believe bold-font, top-of-page warnings about AI content would dissuade some passersby.