If you can’t feel it, the music ain’t vibing. Artificial intelligence has created a dilemma for musicians. On the one hand, it threatens the livelihoods of millions of artistes; on the other, it can turn anyone into a “prompt musician”.

Earlier this year, The Velvet Sundown — with the albums Floating On Echoes and Dust And Silence under their belt — had a good run on Spotify. They soon had more than one million streams on Spotify in a matter of weeks. The “band” went viral and even attracted media coverage from the likes of The Rolling Stone.

Such was the popularity that their music began popping up on popular playlists, creating a domino effect of streams. Then poof…their elaborate music heist was suddenly over for “vocalist and mellotron sorcerer Gabe Farrow, guitarist Lennie West, bassist-synth alchemist Milo Rains, and free-spirited percussionist Orion ‘Rio’ Del Mar.” Even their photograph on the platform was AI-generated. It’s one of the strangest, wildest, most rollicking AI dramas of the year.

The lesson? The need for human curation on streaming services. It’s not going to stop Spotify CEO Daniel Ek from pouring more money into AI tools and tech, even though it’s getting harder by the day to stream music, knowing what’s happening behind the curtain.

More clarity is needed, something the French streaming platform Deezer offered earlier in the year. The service said up to seven out of 10 streams of AI-generated music on Deezer are fraudulent.
“As long as there is money [in fraudulent streaming] there will be efforts, unfortunately, to try to profit from it,” said Thibault Roucou, director of royalties and reporting. “That’s why we’re investing in fighting it — because it’s not going away and we need to be one step ahead.”
According to The Guardian, Deezer deploys a tool it says can detect 100 per cent AI-made content
from the most prolific models.

Making compromises, um, deals

Copyright is on every musician’s lips while the tech industry, to a large extent, has been using publicly available, copyright-protected data to build artificial-intelligence tools. Some labels are brokering deals to stay ahead of the curve.

For instance, Universal Music Group said in late October it had reached licensing agreements with AI music start-up Udio, settling a lawsuit that accused the company of using copyrighted music to train its AI. In short, users of Udio’s AI can create music — including voices and instruments — from text prompts. Udio has agreed with UMG to launch a new platform next year, trained on “authorised and licensed music” and will allow users to customise, stream and share songs.

“These new agreements with Udio demonstrate our commitment to do what’s right by our artistes and songwriters, whether that means embracing new technologies, developing new business models, diversifying revenue streams or beyond,” said Lucian Grainge, UMG’s chairman and chief executive, in a statement.

Udio was earlier accused of using hits like The Temptations’ My Girl to create a similar melody called Sunshine Melody (UMG owns My Girl).

Founded in 2024 by former Google DeepMind employees, Udio’s backers include music artiste will.i.am, Instagram co-founder and Anthropic’s chief product officer Mike Krieger, and venture-capital firm Andreessen Horowitz.

Stockholm-based Spotify, for its part, said in a recent statement: “Some voices in the tech industry believe copyright should be abolished… We don’t. Musicians’ rights matter. Copyright is essential. If the music industry doesn’t lead in this moment, AI-powered innovation will happen elsewhere — without rights, consent or compensation.”

Spotify, too, is taking measures, having removed 75 million “spammy” tracks from the platform in the past year.

Both sides

Some say chatbots and AI models are making music more accessible, like they did for British AI musician Oliver McCann, who goes by imoliver. From indie-pop to electro-soul to country-rap, he can do it all.
His confession to AP is simple: “I have no musical talent at all… I can’t sing, I can’t play instruments, and I have no musical background.”

The 37-year-old visual designer began experimenting with AI to see if it could “bring some of my lyrics to life.” A few months later, he signed with independent label Hallwood Media after one of his tracks hit three million streams.

Tech experts believe generative AI will transform music, but few talk about its effect on the $29.6 billion global recorded-music market, about $20 billion from streaming.

In truth, few people are listening to AI music. Deezer’s estimates earlier this year showed that 18 per cent of songs uploaded daily were purely AI-generated, though they account for only a tiny fraction of streams.

Decades ago, McCartney and Lennon inspired each other across a table. McCann, meanwhile, prompts and re-prompts AI to create up to a hundred versions of a song before he’s satisfied.

Turn the beat around

All musicians — pop to classical — know the power and pitfalls of prompts. Take sarod virtuoso and composer Soumik Datta.

“We already have AI artistes and entirely AI-made tracks. But I’m warmed to see it still can’t replicate the complexity of the sarod, let alone Indian classical music,” he told t2oS. “Tracks as we understand them — with A lines, B lines, choruses and bridges — are easy for AI to replicate. So it pushes me to explore unorthodox structures that break and reinvent form.”

He said this is the moment to “abandon formulas and embrace experimentation.”
“I want to hear long pauses, sudden shifts in melody, grooves that pulse in beats of five or seven-and-a-half. Let’s untangle from Western tempered tuning and find other tonal centres. The harmonic dissonance of traffic or the chaos of wind through a tree are points of deep inspiration. There are rhythms all around us. While AI gets better at making more of the same, this is our chance to feel free of formulas and enjoy the risk of something new.”

For jazz musician Allysha Joy, AI platforms like Udio and Suno have no meaning. Music, she told this newspaper, “is art to express my experiences, to process the world and find connection.”
“It’s therapy for me,” she added. “So I’m not phased or changed in my creative process by these things — and I think that’s true for many artistes.… I feel incredibly sad about AI’s environmental, social and financial impact. It’s pulling people away from their creativity, their connection to themselves and each other, in a culture obsessed with optimisation and productivity. I hate the space it’s taking in music — but it doesn’t change my expression.”

Holly Herndon, a Berlin-based American composer, takes an open approach to generative AI. Her music is built in Max, a visual programming language for custom instruments and vocals. Similarly, Taryn Southern made her album I Am AI using several AI-based tools.

But AI is spiralling out of control. In October, Los Angeles musician Luke Temple, once frontman of indie band Here We Go Magic, learnt that the group had “released” a new track — Water Spring Mountain — with vocals that didn’t sound like his. It was live on Spotify, Tidal, and YouTube. Hello, AI tricksters.

Earlier this year, an AI-generated track appeared on Uncle Tupelo’s page — Wilco singer Jeff Tweedy’s former band. It also happened to electro-pop artist Sophie, who died in 2021, and country singer Blaze Foley, who died in 1989, whose Spotify pages were vandalised with AI songs.

Lessons can be picked up from the career of Prince. He always did what he wanted to do, without surrendering his music to the Internet. When he was alive, he did not hesitate to withdraw his music from all streaming services (except Tidal) without any explanation. “We made money [online] before piracy was real crazy,” he told The Guardian in 2011. “Nobody’s making money now except phone companies, Apple and Google…. It’s like the gold rush out there. Or a carjacking. There are no boundaries…. I personally can’t stand digital music. You’re getting sound in bits. It affects a different place in your brain. When you play it back, you can’t feel anything. We’re analogue people, not digital.”

Welcome to the new Wild West, where AI tool-makers have the guns and the ageing sheriff is fumbling for a plan.