Users are encountering AI-generated songs by artists that don’t exist—many of which are being aggressively promoted by the platform’s algorithm
One day, as Adam Faze listened to Spotify radio, he noticed the same song was coming up over and over again, even as he opted to skip it. It was annoying, but he shrugged it off—at least until he realized that, each time he encountered the track, it was listed under a different artist, with a different title: “Isabelle Morninglocks,” The Brave Android,” “Crash Tortoise,” “Queezpoor,” or “Viper Beelzebub,” to name a few. All of these songs—or song, if you will—were released between now and last August, featuring generic stock images as album art: photos of people crowded around a computer, or an AI-generated man smoking a cigar, or a grassy meadow with a cloudless sky. The worst part? “It’s a song I wish never existed in the first place,” Faze sighs. “It’s the worst song.”
Faze compiled all 50-some versions into a single playlist, and turned to the internet for answers. “I’M LOSING MY MIND. TELL ME AN AI DIDN’T MAKE THIS SONG AND RELEASE IT 100 TIMES,” he wrote in a now-viral Twitter thread.
He’s not the only one having this experience. Recently, a music writer by the name of Jaime Brooks was recommended a song on Spotify: a 46-second instrumental composition attributed to a musician who was also named—get this—Jaime Brooks. “It would be misleading to describe the piece as music,” writes Brooks—the real one—in her Substack, comparing its sound to that of many violin performances, chopped up and re-arranged at random. And yet, despite releasing songs that are fundamentally unlistenable, her musical alter ego has managed to attract a following of nearly 700 monthly listeners—because, as Spotify’s ‘discovered on’ section shows, it was recommended to those seeking ‘classical music for babies,’ the platform’s least-discerning users.
Such playlists have become increasingly popular in the streaming era—because when one need not pay for a specific piece of content to consume it, users are more inclined to turn to pre-existing playlists to soundtrack many of life’s everyday activities. For instance, sessions in which the user is listening to pre-curated playlists account for a vast majority of streams on Amazon Music—and mood music, as a whole, represents one of the industry’s fastest-growing sectors.
“It’s been said that the proliferation of AI tools will democratize the music industry—but if the slew of fake artists on Spotify is any indication, it’s equally likely to result in a flood of algorithmically-optimized slush.”
Today’s increasingly passive audience—many of whom might be eating, exercising, or even sleeping while they listen to pre-curated playlists—helps explain how these scams might manifest: By generating harmonies that resemble those of classical music, and deploying familiar sounds in a relatively nonsensical fashion, “artists” like Jaime Brooks are able to bamboozle the algorithm into serving their work up alongside that of real composers, while simultaneously bypassing the copyright filters designed to block infringement of existing work. “It really does feel like the internet, as a whole, is just gonna become a cesspool of junk very quickly,” says Faze dispiritedly. “I’m like, Is this the dropshipping of music?”
The impact of AI on the music industry has already been hotly debated this week, with viral deepfakes being yanked off streaming platforms after Universal Music Group urged Spotify to obstruct AI companies from training their programs on copyrighted work, lest they harm the income of living artists. “[This] begs the question as to which side of history all stakeholders in the music ecosystem want to be on: the side of artists, fans and human creative expression, or on the side of deepfakes, fraud and denying artists their due compensation,” stated a spokesperson from the company, an industry titan which counts Drake among its many signed artists.
But as it turns out, UMG has another horse in this race: because while they may not be releasing fake tracks by popular artists, it’s among the many major labels and streaming platforms that have allegedly been populating playlists with “fake artists” for years—many of whom are paid a set fee to compose songs, which are then listed under the names of several artists, with Spotify pocketing the full profits. At the time of writing, UMG and Spotify have not responded to requests for comment, and the former has previously denied claims that they were playlisting fake artists—but that doesn’t explain why so many of the “chill vibes” they’ve curated appear to be created by pseudonymous musicians with no digital footprint, or why they’re being so aggressively promoted to users like Faze.
It’s been said that the proliferation of AI tools will democratize the music industry—but if the slew of fake artists on Spotify is any indication, it’s equally likely to result in a flood of algorithmically-optimized slush that makes it harder for users to access the kind of art they value. It begs the question: What responsibility do streaming platforms have to convey how songs are made—and should consumers be paying more attention to what we’re listening to, lest the future of music be determined while we’re asleep at the wheel?