6 days ago
You could be listening to fake artists on Spotify and not even know it
Not only has there been a marked increase in AI-generated music released by fake artists, but music streaming services are also pushing those artists into your playlists.
The issue of fake or bot accounts on platforms like Spotify has again made headlines this week, thanks to Spotify 'verified' band The Velvet Sundown being unmasked as an AI imposter.
The '70s-inspired psych-rock band clocks over 858K monthly listeners on Spotify alone and is being recommended in subscribers' algorithmic playlists, like Discover Weekly.
However, it was its AI-tinted promotional images, fabricated band bio including the nebulous descriptor 'like a scent that suddenly takes you back somewhere you didn't expect', that gave them away — not to mention a lack of live show dates, a sparse digital footprint and the tell-tale, frequency-glossed sheen of algorithmic production.
When a spokesperson from the project spoke to Rolling Stone, they confirmed the band used AI music generator platform Suno, and called the entire project a 'marketing stunt'.
'It's trolling,' said the spokesperson, who used the pseudonym Andrew Frelon for the interview.
'People didn't care before, but now we're in Rolling Stone,' he said.
As previously reported by the issue of fake streams is diverting royalties away from real human performers; estimated to take about $US2 billion out of artists' royalties per year.
But now, it's getting more difficult to decipher what's fake and what's not.
Even 50 Cent is enjoying AI music, and is even sharing it to his millions of followers.
Sharing a clip of himself singing along to AI hip hop artist Nick Hustles, the rapper can be seen mouthing the lyrics to the Nick Hustles song Why U N ** s Gotta Hate.
Loading Instagram Post
Nick Hustles has over 238,000 monthly listeners on Spotify, and even has his own official 'This is Nick Hustles' playlist (the same treatment given to real-world artists), and has been featured in the platform's Release Radar playlists.
The liner notes for Nick Hustles outs the project as AI-driven.
As reported by MBW, Hustles is the creation of Pennsylvania-based Nick Arter, a self-described AI Storyteller and founder of AI For The Culture, an Instagram page with over 100,000 followers dedicated to fictional AI artists posing as the forgotten voices of 60s, 70s and 80s soul and funk.
Locally, Australia's industry has come out in droves to push for regulation and transparency from generative AI models.
Dean Ormston, CEO of music body APRA AMCOS, said we need clear, enforceable rules around the training of AI models.
'Without this, Australian and New Zealand creators risk being exploited, their works devalued, and their voices erased,' he said.
Sharing excerpts from an article published by MBW to Facebook, songwriter and media personality Mark Holden said, 'As if artists, songwriters and musicians aren't facing enough in the digital economy, then there's this.'
Victoria-born UK-based musician Nick Cave said, 'That's (AI music's) intent: to completely sidestep the inconvenience of the artistic struggle (by) going straight to the commodity.
'Which reflects on us, as what we are as human beings: things that consume stuff.
'We don't make things anymore; we just consume stuff.'
Warrnambool artist Didirri took to Instagram to share his frustration with the influx of AI music.
'This AI sh*t that is putting slop onto the Internet and generative, iterative, crappy art…,' he paused.
'I just kind of have to stop and so I'm going full luddite, and I'm going to start a physical mailing list.'
Didirri pushed his fans to attend more live shows if they shared his frustrations.
'If you're feeling existential and weird about the world right now and how fake it's getting, maybe go hang out with some human beings in a room that make art,' he said.
Australian producer Tushar Apte — who recently released the single She's Miss Cali with Idris Elba and Snoop Dogg with his The Dualists project — has what he's calling a 'short-term' tactic for artists frustrated by AI-generated music infiltrating streaming platforms.
His suggestion? Flood the system, in return, by uploading AI-generated tracks under the fake band's name.
This will make it difficult for the original AI creators to maintain control, and redirect any royalties to a good cause.
Taking to Instagram he said, 'If you're an artist, everybody knows that if someone fraudulently uploads a song to Spotify with your artist's name on it, it's pretty hard to take that thing down, as a featured artist or whatever it is.
'So ... You could make a song on Suno, or whatever, upload it and say it's by this band.
'And it'll be hard for them to take that down. Do a bunch of them. Do 50 of them. Do 100 of them.
'Collect the money from that. Let's put it back into a charity or into a fund that helps songwriters.'
How popular are fake artists on music streaming services?
A report by MBW detailed 13 AI-made 'artists' currently active on Spotify, with approximately 4.1 million cumulative monthly listeners between them.
AI-generated country singer Aventhis boasts more than one million monthly listeners and his song Mercy On My Grave has already surpassed two million streams.
However, Aventhis doesn't exist in any human sense.
His voice, image and entire persona were created using AI tools, likely through platforms like Suno or Udio.
MBW noted that the human songwriter behind Aventhis is David Vieira, who admitted in online comments that the artist's voice and image were entirely AI-made, with Vieira providing only the lyrics.
In the space of just four months, Aventhis has released three full albums, including 57 tracks, all available across major streaming services.
Industry experts warn the scale of its reach is now impossible to ignore with fake acts quietly dominating algorithm-driven recommendations on services like Spotify, Apple Music and YouTube Music.
APRA AMCOS' Dean Ormston said the onus is on streaming platforms to provide transparency and safeguards.
'Our concern regarding Generative AI and the Large Language Models (LLMs) like ChatGPT arises from the complete lack of transparency platforms have demonstrated in terms of acknowledging the music, art, books and screen content that has been 'scraped', 'mined', 'listened to', 'trained on', or to use a better word, copied, to create their outputs,' he said.
