Photo Credit: Mohamed Nohassi
An band’s music was stolen, pitch-shifted, and re-uploaded under a fake name with AI-generated artwork, snatching up over 90% of their Spotify royalties.
Musician Owen Lyman-Schmidt found out that his work had been stolen when a longtime fan of his band, Makeshift Hammer, messaged him to let him know an album on Spotify sounded as if someone had taken his music and “distorted it a bit.” They sent a link to an album called Blue Road by an artist named Carey Dupont.
“I had never heard of Carey Dupont. And if you looked closely, the album art didn’t quite make sense,” Lyman-Schmidt wrote. “Dupont had no other releases that I could find and was seemingly otherwise unknown to the internet.”
Lyman-Schmidt describes Makeshift Hammer as “Philly’s premier mandolin-bass guitar-junk percussion gutter-folk duo, which is a fancy way to say we know a thing or two about being unknown.” He and his bandmate Bobby have been playing together “for a dozen years,” and music isn’t their day job; in fact, Lyman-Schmidt is a private detective.
“But Carey Dupont seemed even more off the grid than we were. They had no website, no social media, no dead ticket links to past shows, no profile in alt-weeklies,” he explained. “Nothing except Blue Road, the album, on every streaming service.”
The names of the songs, like the album art, didn’t make much sense either—but they were all similarly named to existing Makeshift Hammer tracks. “All My Friend” instead of “All My Friends,” “If Not Obvious” instead of “If It’s Not Obvious,” and “Banker and a Liar” instead of “Bankers and Liars.”
And the music itself was the most egregious part. It was the actual songs Makeshift Hammer had recorded, but either sped up or slowed down, which distorted the sound but still left them easily recognizable. Of course, Lyman-Schmidt immediately recognized the digitally altered sound of his own voice.
“The fear that Carey Dupont would replace us wasn’t just some abstract notion; in the year before we found it, the tracks on Blue Road had been listened to close to 50,000 times each,” wrote Lyman-Schmidt. “By contrast, many of the original recordings had only 1,000 to 2,000 listens despite being released four years earlier. Someone was using our music to play the streaming game and was massively outperforming us.”
Makeshift Hammer’s fans and friends alike agreed that there must be some way to address the issue. “The theft is so blatant,” wrote one listener. Soon, the duo learned that another local Philly musician, Katie Feeney, better known as Roberta Faceplant, had experienced a similar problem.
“I’m still trying to get [the copies taken] down,” she told them.
In Makeshift Hammer’s case, those behind Carey Dupont’s album took their existing recordings, altered them “the bare minimum” to avoid easy detection by music identification software, and repackaged them before “cranking out the automated listens.”
By now, the tracks from the Blue Road album had surpassed 650,000 listens, and it seemed near-constant—obviously inflated listens using bots. The fake listener bots have, as Digital Music News has covered extensively, been plaguing Spotify and other music streaming platforms well before the birth of today’s AI.
“Online royalty estimators suggest that number of listens would pay out around $1,600 to $2,600. Not a huge heist, but rinse and repeat over a few hundred low-profile bands and you have a living wage, without the 10-hour drives between gigs and venue merch cuts faced by actual musicians.”
While Spotify would have you believe that its payout model is artist-centric, ask any musician who doesn’t have a song on the Spotify Billions Club playlist how accurate that is. In Spotify’s own verbiage on its website, “artists get paid based on their share of total streams on Spotify. If your music accounts for 1% of all streams, you get roughly 1% of the royalty pool.” Which is bad news for small artists, and even worse news for artists competing with their own stolen music.
For Spotify, the same amount of money is paid out, regardless. Just because Carey Dupont is getting a bigger piece of the pie than Makeshift Hammer doesn’t mean Spotify is making that pie any bigger. It just means Makeshift Hammer is having to share their piece with someone who stole their work.
“This royalty structure allows Spotify to pass off the true costs of scammers to artists rather than pay for its inability or unwillingness to monitor its own platform,” said Lyman-Schmidt. “Ducking the real financial consequences of the model is apparently how Spotify can afford to offer most of the recorded music in the world for immediate consumption at any time at the suspiciously reasonable price of $12 a month.”
But if the Carey Dupont situation wasn’t bad enough, Makeshift Hammer had the same thing happen about six months later, this time in the form of an album called Powerful Thinking of Hayden Donne.
Again, this artist had no other music online, no online presence, and obviously AI-generated album art. Again, the songs were altered just enough not to be auto-detected, and named only slightly differently from their source. However, this time only half of the music was Makeshift Hammer’s, mixed together with a pop artist’s work in a strange clash of genres.
Again, hundreds of thousands of streams were generated by this new album before it suddenly disappeared. Lyman-Schmidt was afraid if he just reported these artists to Spotify, they would simply disappear and nothing else would come of it. So instead, he did some digging to find the distributor of Carey Dupont and Hayden Donne’s albums.
Ultimately, he discovered that both scam artists were distributed via SoundCloud. He reached out to SoundCloud about the matter and received an automated message asking for more information. He also reached out to Spotify and received “nothing but silence.”
While the majors ink deals with AI-generated music platforms, independent artists like Makeshift Hammer watch their already small piece of the streaming royalties pie get cut down to a sliver. Bad actors steal their work and pass it off as their own with minimal alterations and an army of bot-listeners. And that was already a problem before the onslaught of generative AI.
The combination of AI music fraud and artificially inflated streams is a symptom of a larger streaming-era problem: platforms have made music far easier to distribute at the cost of authenticity. Distorted streams, royalties, and artist visibility are plaguing smaller artists who already have to fight to be seen and heard. So what are the major industry players willing to do about it?
