The AI Band That Sparked Spotify’s Content Regulations Row

A fake band created entirely by AI has racked up millions of streams on Spotify, sparking a fierce debate about how music platforms should handle synthetic content.

The Velvet Sundown appeared on streaming services in June, presenting itself as a regular folk rock band with polished photos and a carefully crafted sound.

Within weeks, the group had notched up millions of listens on Spotify.

But music fans quickly started noticing something odd about the whole setup.

The band’s promotional photos had that slightly unsettling quality that’s become the hallmark of AI-generated images.

Things got worse when someone claiming to be connected to the project revealed that the band had used Suno, an AI platform for creating music, to generate their tracks.

The team behind The Velvet Sundown initially tried to deny these claims through social media, but eventually came clean about the artificial nature of the project.

The admission has triggered a backlash from across the music industry and raised awkward questions about how streaming platforms police their content.

Music industry demanding transparency from AI companies

Ed Newton Rex runs Fairly Trained, a non-profit focused on AI ethics and he thinks the incident is exactly what musicians have been fretting about.

“This is exactly what artists have been worried about, it’s theft dressed up as competition,” he says.

“AI companies steal artists’ work to build their products, then flood the market with knock-offs, meaning less money goes to human musicians.”

The controversy has exposed how little oversight there is of AI-generated content in music streaming.

Current rules don’t require platforms to flag synthetic music, leaving both artists and listeners in the dark about what they’re hearing.

Roberto Neri, who runs the Ivors Academy, a trade body representing music creators, argues: “AI-generated bands like Velvet Sundown that are reaching big audiences without involving human creators raise serious concerns around transparency, authorship and consent.”

Sophie Jones, Chief Strategy Officer at the British Phonographic Industry (BPI), which represents the UK’s recorded music business, says: “We believe that AI should be used to serve human creativity, not supplant it.”

The BPI has been pushing for more transparency from AI companies about how they train their systems and generate content – showing the growing worry across the industry that the current free-for-all approach isn’t working for musicians trying to make a living.

Writer Liz Pelly has spent years documenting how AI models get trained using music without artists’ permission or payment. In her book “Mood Machine: The Rise of Spotify and the Costs of the Perfect Playlist,” she warns that AI-generated music could make everything sound the same.

“Musical trends produced in the streaming era are inherently connected to attention, whether it’s hard-and-fast attention-grabbing hooks, pop drops and chorus loops engineered for the pleasure centres of our brains, or music that strategically requires no attention at all - the background music, the emotional wallpaper, the chill-pop-sad-vibe-playlist fodder,” she explains.

Streaming platforms show fragmented responses to AI content

Different streaming services have taken different approaches to handling AI-generated music, highlighting the lack of industry-wide standards.

Deezer, the French streaming service, has actually done something about it, by rolling out detection software that spots AI-generated tracks and tags them for users. 

Aurélien Hérault, Deezer’s CIO, sees this as a transparency issue: “For the moment, I think platforms need to be transparent and try to inform users.”

He suggests that current tagging might just be a temporary fix while AI becomes more embedded in how music gets made.

Spotify has taken a more hands-off approach, insisting it doesn’t favour AI-generated music over human-made tracks. 

A spokesperson says “all music on Spotify, including AI-generated music, is created, owned and uploaded by licensed third parties.”

But Spotify has been here before. Liz’s research proves how the platform stuffs its playlists – especially instrumental ones – with AI-generated songs to avoid paying artist royalties. 

The strategy helps boost the company’s margins by reducing the money it pays out to musicians.

The Velvet Sundown found its biggest audience on Spotify, which has led some industry watchers to wonder whether the platform actively promoted the synthetic content. 

The exact mechanics of how the band’s music gained traction remain murky.

Most streaming platforms currently treat AI-generated music the same as human-created content, as long as it’s uploaded by licensed parties. 

Unsurprisingly, this approach has created uncertainty about how platforms will handle future cases of synthetic content that attracts large audiences.