The image depicts not a real artist, but AI-avatars called "Xania Monet", created by Telisha Jones through artificial intelligence. Photo: Instagram, @ xania_monet 

Xania Monet has around 460,000 monthly listeners on Spotify and has just - after a bidding war - been "signed" for three million dollars. But Xania Monet doesn't exist in physical reality. She was created by American Telisha Jones, who uses AI to generate the music and who signed the record deal with Hallwood Media in 2025.  

At det overhovedet er muligt at indgå aftaler med en digital efterligning af et menneske, møder af skarp kritik. Sangerinden Kehlani sagde det klart på TikTok:  

"There's an AI R&B artist who just signed a multimillion-dollar deal... but they're not doing any of the work."  

For Sara Indrio, the problem is not only artistically, but also fundamentally ethically indefensible and requires attention in everything from the way we legislate to the way we talk about it.  

"We should be careful about even using the term 'artist' for computer-generated products. They have nothing to do with a human, and services should at least disclose when something is created by AI and is an imitation of a real human," she says.   

She points out that while it may be fun and exciting right now that technology is opening up new possibilities, we must remember that it is other people's experience, professionalism and humanity that it is all based on. New generations should have access to the same originality and authenticity that the rest of us grew up with: 
 
"AI can perform without fatigue and become the sum total of everything the rest of us strive for in a matter of minutes, but it cannot be human. It cannot feel, fail or change through life - and that is the essence of art. Also the art that technology and music services are currently exploiting."     

A new type of exploitation of the streaming economy  

The exploitation of artists in the business model of streaming platforms is nothing new. Since the mid-2010s, so-called streaming farms mass-produced music and used bots to generate royalties with fake streams. Money taken from real rights holders.  

Journalist Liz Pelly documented back in 2016 that Spotify itself used so-called "ghost artists" on functional playlists for sleep, focus and relaxation. The music was produced by hired musicians, but royalties went to the platforms themselves.  

With AI, the barrier is lowered even further:  

Now anyone can mass produce music in seconds - even people with no musical background, like Telisha Jones - and profit from it in a system that is becoming less favourable to music artists.  

This development is reflected in the fact that music created by artificial intelligence is being uploaded to streaming services en masse. According to music service Deezer, the service receives 50,000 tracks a day that are AI-generated, according to Music Business Worldwide. 

AI exploits artists' works - without permission  

AI systems are trained on huge amounts of copyrighted material, often without consent or payment to the artists. It's a practice both artists and lawyers warn against.  

Sara Indrio put it succinctly at the EU Presidency AI Conference:  

"With AI, we're not just protecting our artistic ideas - we're fighting to protect who we are and what our bodies produce. This is new and unprecedented."  

She points out that artists are already experiencing pressure to give up rights to AI training in contracts - a trend that the Danish Artist Association is trying to counteract in the negotiation of music contracts. 

Expert group: AI violates the purpose of the law  

The Ministry of Culture's Expert Group on Copyright and Artificial Intelligence supports this assessment. According to the expert group, training commercial AI systems on protected material without authorisation is contrary to the purpose of the Copyright Act and ethical practice.   

The group emphasises that content created by artificial intelligence can directly compete with - and at worst outperform - human-created contentand therefore recommends legal requirements for transparency in training data, a ban on digital imitations of voices and appearance, and technical safeguards to prevent illegal uploading of copyrighted material.  

"Action is needed now. The first commercial AI systems have fundamentally violated copyright," the report concludes.
 
Read Recommendations from the Expert Group on Copyright and Artificial Intelligence here.   

AI music could cost the Danish music industry billions  

The economic consequences also look serious. A new analysis from Koda and IFPI Denmark, prepared by HBS Economics, shows that "In the absence of political intervention, the Danish music industry risks losing DKK 6.9bn by 2030 as a result of fully AI-generated music".  

By 2030 alone, according to the report, the loss is estimated at around DKK 2bn per year - equivalent to 28 % of the industry's revenue from recorded music. This is mainly due to the fact that AI-generated music can replace human-created music on streaming services, social media and in background music.  

New acquisitions and industry shifts 

That relationship gets really 'tricky' when AI services team up with record producers. Recently, an agreement was signed between Universal Music Group (UMG) and AI service UDIO, which will launch a new streaming service next year. This is of great concern to the International Association of Artists' Organisations (IAO).  

"If I think about what could happen with payment limits on streaming services with an even stronger Universal, I definitely wake up in a cold sweat," says chairman Nacho Vega, who urges the EU to block the deal in view of the market dominance UMG will gain.  

"It's very worrying that big AI deals are now being made with as little transparency as the deals the big companies previously made with streaming services." 

You can read the full letter from Nacho Vega here.

Også Warner Music Group har netop indgået et partnerskab med SUNO, som forliger den igangværende retssag mellem de to parter.  

The agreement includes SUNO being able to use Warner Music Group signed artists' voices, faces and works - but not without permission from the artist themselves, says Robert Kyncl, CEO of WMG, to Music Business Worldwide. 

However, it is still unclear exactly what such partnerships and agreements include, especially which rights are used and how, says Jens Skov Thomsen, Head of Law at the Danish Artist Association:  

"The agreement between major labels and AI companies is apparently enough to drop otherwise ongoing lawsuits. But what's in it for the artists when label and AI service shake hands? And are artists' interests adequately protected in such a process? Do Universal and Warner Music Group give access to train AI on their catalogue of music releases? The lack of transparency and uncertainty about what is being traded between the majors and the AI companies is worrying."

The Danish Artist Association believes that only the artists themselves can give consent for AI training based on their recordings and sound. Even if the rights to certain recordings have been transferred to a producer.  

"Manufacturers simply do not have the necessary rights to enter into licences and agreements on training and other uses in an AI context. That's why it's now even more important to read the small print before signing contracts in the future," says Jens Skov Thomsen.   

Legal battles in several countries  

Courts in several countries are now considering how copyright should apply in the age of AI. This year, a German court ruled that OpenAI infringed copyright by storing and reproducing protected song lyrics from GEMA's catalogue. The judgement is the first of its kind in Europe and orders OpenAI to pay compensation and delete the material.  

In Denmark, Koda has sued AI company Suno for illegal use of Danish music, including works by MØ, Christopher and Aqua, in the training of their music generator.  

Some artists are experimenting with AI as a tool. Music producer Timbaland, for example, has launched AI artist TaTa via his company Stage Zero. But the project is based on tools developed by Suno - which is currently facing legal claims for illegal training.  

Sara Indrio, chairperson of the Danish Artist Association, emphasises how important it is to get the legislation in place:  

"Working with the law will make it easier for us to ensure that artists do not unwittingly consent to their works or voices being used for AI training or digital impersonations. We encourage all artists to seek legal advice and we are now introducing AI caveats in all new contracts and agreements." 

Do you have questions about your contract? Contact the Danish Artist Union's legal department at jura@artisten.dk and get advice.

 

Dansk Artist Forbund also works internationally to protect artists through the global federation of artists' unions - FIA  - International Federation of The Arts. See the FIA's new guide on AI here. 

Major Labels is now partnering with AI services 

Both Warner Music Group and Universal Music Group, who have previously sued the AI services for theft of rights related to AI training, are now entering into licence and partnership agreements with the same services. Most recently, Warner Music Group has just entered into a partnership with SUNO, which settles the ongoing lawsuit between the two parties.  

The agreement includes SUNO being able to use Warner Music Group signed artists' voices, faces and works - but not without permission from the artist themselves, says Robert Kyncl, CEO of WMG, to Music Business Worldwide. 

However, it is still unclear exactly what such partnerships and agreements include, especially which rights are used and how, says Jens Skov Thomsen, Head of Law at the Danish Artist Association:  

"The agreement between major labels and AI companies is apparently enough to drop otherwise ongoing lawsuits. But what's in it for the artists when label and AI service shake hands? And are artists' interests adequately protected in such a process? Do Universal and Warner Music Group give access to train AI on their catalogue of music releases? The lack of transparency and uncertainty about what is being traded between the majors and the AI companies is worrying."

The Danish Artist Association believes that only the artists themselves can give consent for AI training based on their recordings and sound. Even if the rights to certain recordings have been transferred to a producer.  

"Manufacturers simply do not have the necessary rights to enter into licences and agreements on training and other uses in an AI context. That's why it's now even more important to read the small print before signing contracts in the future," says Jens Skov Thomsen.   

Do you have questions about your contract? Contact the Danish Artist Union's legal department at jura@artisten.dk and get advice.