This was the year that AI-generated music went from jokey curiosity to mainstream force. Velvet Sundown, a wholly AI act, generated millions of streams; AI-created tracks topped Spotify’s viral chart and one of the US Billboard country charts; AI “artist” Xania Monet “signed” a record deal. BBC Introducing is usually a platform for flesh-and-blood artists trying to make it big, but an AI-generated song by Papi Lamour was recently played on the West Midlands show. And jumping up the UK Top 20 this month is I Run, a track by dance act Haven, who have been accused of using AI to imitate British vocalist Jorja Smith (Haven claim they simply asked the AI for “soulful vocal samples”, and did not respond to an earlier request to comment).
The worry is that AI will eventually absorb all creative works in history and spew out endless slop that will replace human-made art and drive artists into penury. Those worries are being deepened by how the major labels, once fearful of the technology, are now embracing it – and heralding a future in which ordinary listeners have a hand in co-creating music with their favourite musicians.
AI music platforms analyse huge amounts of recorded music in order to learn its sounds, structures and expressions, and then allow users to create their own AI-generated music via text or speech prompts. You might ask for a moody R&B song about a breakup sung by a female vocalist, and it will come up with a decent approximation of one, because it’s absorbed hundreds of such songs.
Artists and labels initially saw AI as the biggest existential threat since Napster-fuelled piracy: if not a replacement for human creativity, then certainly a force that could undermine its value. Gregor Pryor, a managing partner at legal firm Reed Smith, says background music for things such as advertising, films and video games, where you’re not relating to a personality as you would in pop music, “is where the real damage will be done” first of all. “People will ask: why would I pay anyone to compose anything?”
Aware of the scale of the shift, last year the Recording Industry Association of America, representing the three major labels, initiated legal action against AI music companies Suno and Udio for copyright infringement, alleging they had trained their AI platforms on the labels’ artists without their permission. But then there was an extraordinary about-turn. They didn’t just settle the matter out of court – Universal Music Group (UMG) then partnered with Udio, and Warner Music Group (WMG) with Udio and Suno. They also have deals in place with AI company Klay, the first to get all three major labels on board, adding Sony Music (discussions with indie labels are ongoing). WMG chief executive Robert Kyncl has said these recent deals are to ensure the “protection of the rights of our artists and songwriters” and to fuel “new creative and commercial possibilities” for them, while UMG chief Lucien Grainge heralded “a healthy commercial AI ecosystem in which artists, songwriters, music companies and technology companies can all flourish and create incredible experiences for fans”.
Kyncl made another bold statement as to why these deals are taking place: “Now, we are entering the next phase of innovation. The democratisation of music creation.”
Announcing its Universal tie-in, Udio chief executive Andrew Sanchez has said Udio users will be able to “create [music] with an artist’s voice and style”: so not just create the aforementioned moody R&B song, but one with a specific existing artist’s voice. He also says Udio will allow users to “remix and reimagine your favourite songs with AI … take your favourite artists, songs or styles and combine them in novel ways. In our internal experimentation, the team has gotten some truly remarkable and unusual results that will definitely delight.”
Klay meanwhile states that “fans can mould their musical journeys in new ways”, but it’s essentially the same offering: a subscription service where you can manipulate the music of others, or create your own from it. Ary Attie, Klay’s founder and chief executive, says his company will properly compensate artists whose work is used, and won’t supplant the work of human musicians: “This technology is not going to change any of that.”
Klay is a rarity in that it signed up all three major labels before it started training its AI system on their music: “A core part of our philosophy,” Attie says. He argues that rival AI companies – he doesn’t name names – have been “acting in a way that doesn’t respect the work of artists, and then being forced into a corner”. Suno did not respond to an interview request; Udio claimed its executives were “extremely swamped” and therefore unable to answer questions. The current, and synchronised, messaging from labels and gen AI companies with licensing deals is that they all respect both art and artists and that their deals will reflect this.
They are also positioning gen AI as the single biggest democratising leap ever in remix culture, effectively enabling everyone to become musically creative. The counterargument is that, by lowering all barriers to entry and by allowing the manipulation of a song or a musician’s character at scale, it vastly devalues and negates the creative act itself.
***
But what do musicians actually think of the prospect of their work being used to train AI, and reworked by the general public? “Everybody should be selling or licensing their voice and their skills to these companies,” Dave Stewart of Eurythmics argued to me this week. “Otherwise they’re just going to take it anyway.” That view is directly countered by the major labels and AI companies, who have insisted artists and songwriters get to opt in to have their music made available, and if they do, get royalties when their music is used to train AI, or manipulated by users on platforms such as Udio, Suno and Klay.
Others take a grimmer view about how these companies might reshape the industry. Irving Azoff, legendarily forthright artist manager and founder of the Music Artists Coalition in the US, responded to the Universal/Udio deal with biting cynicism. “We’ve seen this before – everyone talks about ‘partnership,’ but artists end up on the sidelines with scraps,” he said. In the wake of the same deal, the Council of Music Makers in the UK accused the major labels of “spin” and called for a more robust set of artist-label agreements. And the European Composer and Songwriter Alliance says there is a disturbing “lack of transparency” around the deals (though more detail is likely to emerge on what users can do with any music they create, and any potential commercial uses of it).
Catherine Anne Davies, who records as the Anchoress and also sits on the board of directors at the Featured Artists Coalition (FAC), has many reservations here. “Most people don’t even want their work to be used for training AI,” she says. “I’m on the dystopian side, or maybe what I call the realist side of things. I’m interested in the way that AI can be assistive in the creative process – if it can make us more efficient, if it can streamline our processes. But generative AI for me, in terms of creative output, is a big no-no at the moment. I’m yet to be convinced.”
Musician Imogen Heap feels that AI itself is not to be feared as a tool – she uses an AI she calls Mogen to listen to every aspect of her life, with a view to it being a creative partner (as explored in a recent Guardian article). To help address some of the issues, she has created Auracles, an artist-led, non-profit platform she hopes will be the place where the rights and permissions around AI are set out. It’s not enough to say you’re happy with your music being used by AI, she says – instead, what’s needed are “permissions that grow and evolve over time”.
Other companies are cropping up with similar offers. “We must protect the artists at all costs,” says Sean Power, chief executive of Musical AI, who aims to give musicians “an exact portion of the influence they’re having on all the generative outputs” – meaning compensation every time even a tiny bit of one of their songs is used by a user of Udio et al.
Terms of these deals are undisclosed, but labels are likely to be seeking settlement for any past use of their artists’ copyrights as well as an advance on future use, plus an equity stake in the platform. And while artists will be able to opt out of including their work, they probably won’t be consulted on these partnerships going ahead, with this lack of consultation being something that artist representative bodies such as FAC have been particularly critical of. “The big artists, the labels need to be nice to; those who have a platform will be consulted to some degree,” says a music licensing expert, speaking anonymously. “The very few, who as individual artists are able to make a dent on share price, will have approval.”
I approached Universal, Sony and Warner about the specific concerns raised by artists here: namely limited transparency around the deals, their commercial terms and how opt-ins work; if there is a risk of gen AI undermining existing revenue sources; and if there is significant artist refusal to assign their works for gen AI training. None of the companies would comment on the record about the specifics. Though in an internal Universal memo about AI deals, sent to all staff earlier this year and seen by the Guardian, Grainge said “we will NOT license any model that uses an artist’s voice or generates new songs which incorporate an artist’s existing songs without their consent.”
The Guardian understands that labels are currently having discussions with artists and their managers to better explain how these deals will work and why they believe they can bring in additional revenue, although they will need to convince artists that gen AI will not damage other sources of income, notably from streaming.
But it isn’t clear whether consumers will actually pay to play around with music in the way Udio and others hope they will. AI is the single biggest hype category in Silicon Valley right now, with an average of $2bn of venture capital investment going into AI companies every week in the first half of this year. Sundar Pichai, chief executive of Alphabet (parent company of Google), recently warned of the catastrophic domino effect across the tech sector if this AI bubble bursts, a concern the Bank of England also recently raised.
Reed Smith’s Gregor Pryor argues that AI music could, counterintuitively, end up being positive for human musicians. “By its nature, AI is derivative and cannot create new music,” he says. “Some investors in music catalogues that I speak to say it’s good for artists, because music ‘verified’ as created by humans will have greater value.”
Artists will frame their work as having an invaluable human essence, their music speaking entirely from the heart, but it will become incrementally more difficult for the casual listener to distinguish between music created by a human and that created by AI. The Guardian understands that radio stations and DJs are currently extremely nervous about AI-powered music slipping through their quality filters, effectively hoodwinking them and hanging question marks over how their playlists work. The example of Papi Lamour might force them to do much greater due diligence on what they put forward for airplay consideration. Or they could be the first trickles of a flood that roars through radio and streaming services as the boundaries between AI and human-created music crumble.
Davies is especially worried about artists not thinking through the long-term implications of licensing to AI services. “We cannot think of ourselves selfishly as entities that will be unaffected, because the entire ecosystem will experience a knock-on effect financially. What about your fellow composers and creators? But also what about the generations to come after? Are we fucking this completely, just to make sure that we can pay our mortgages now?”
AI’s current level of sophistication means it is really producing composites of existing music, creating a Frankenstein’s monster of melodies. However, when AGI (artificial general intelligence) finally arrives, with Anthropic co-founder Dario Amodei suggesting that could happen as soon as next year, we will be catapulted into an exhilarating and terrifying realm of uncertainty for the future and the purpose of human-created art.
“It’s literally happening under our noses,” warns Davies. “We should be so much more concerned than we are.”
