The Impending Implications of AI Music

AI is continuously expanding and improving both independently and in its human development. The rise of AI has been controversial from its inception. The main reason for this is the use of AI to replicate existing creations made by humans. The artwork of a machine is always going to feel different from artwork done by a human, because it’s missing one main component: soul. A person pouring themselves into their artwork is going to reflect that emotion, that intensity. AI can try to replicate that, but it’s never going to hold the same soulful connection – it's always going to be hollow.  

As the use of AI rises, it’s slowly become a problem in the creative sphere. Human artists are being replaced by bots to create things like logos, commercials, and even music. And AI music is what we’re going to focus on because the rise of AI music may seem harmless, but like any other reproduction or facsimile of artwork, it’s going to hurt the human creatives who are in the same realm.   

There are benefits to adding AI to the music industry, but they also come with drawbacks. Some of those benefits include the automation of promotional materials and the rise of AI music generators. Allowing an AI to target audience, schedule posts, and generate marketing content can save artists’ time so they can focus more on the music they want to make. There is, however, the concern that AI posting is manipulating fans in a way if they don’t know it’s a computer engaging with them. 

While music generators can be useful tools for musicians to try out how different sounds fuse together, if they’re used as methods of creation instead of tools to create a different product, many feel that the music comes off less authentic and real. AI’s production of something based on a description, at the very least, can get the creative juices flowing. Building off something is a lot easier than creating from nothing–yet sometimes not as satisfying.  

There needs to be a balance in the use of AI to benefit the processes without losing the necessary human component that makes the creation of music so special. 

One aspect that makes the rise of AI music so concerning is how difficult it can be to tell that it’s AI. As the systems improve, it will become harder to distinguish the difference, and that’s a little (very) terrifying. One area where artists are seeing the most hurt from AI music is on Spotify. Musicians already compete with other musicians for listens on the platform that gives meager royalty payments per stream, but allowing AI music on the platform with no regulation adds more competition to the mix. 

Listeners on Spotify began noticing songs getting recommended to them that they’d never heard of yet had hundreds of thousands of streams. Artists like Aria Sai are theorized to be an AI artist. It’s not hard to jump to that conclusion looking at the photos created to represent the “artist”, including an Instagram account, and the fact that three entire albums were supposedly released by Sai in 2024. Aria Sai has about 78.5K monthly listeners, is a verified musician on Spotify, and has supposedly released four albums in the last year and a half. This is just one of many (alleged) AI-created musicians on the platform. 

Some reading this may be wondering, who really cares? Just don’t listen to the AI artists and you’re good...right? Well, given the Spotify shuffle algorithm that automatically kicks on after you finish an album, they can recommend whatever they like unless you directly click that you want the album to repeat. When you first press shuffle on a playlist, it wants to do the “smart shuffle” method first, which includes Spotify recommending songs for the queue based on other songs in your playlist. You have to manually toggle it to regular shuffle to stop the feature. You could be listening to a song created by AI and not even realize it until you look further into this new artist that you’re enjoying a song from.  

But okay, sure, it’s not a big deal. It’s just another method that the platform itself (this is all an alleged theory, don’t sue me) can make money back through its streaming revenue. What’s stopping the platform from creating a fake artist, uploading AI generated music connected to that artist, and then feeding that music into popular playlists?  

And let’s say that Spotify isn’t doing it (maybe they really aren’t, it’s just a theory after all). Anyone can find a distributor to help them upload to Spotify. What’s stopping a regular person from creating AI-generated music, uploading it to Spotify under a nonexistent artist, getting hundreds of thousands of streams, and raking in that revenue? Where’s the ethics in having a computer create music for a human to profit from that they did not put the work into creating? How is it fair for a human being with a passion for music to put their heart and soul into a song, put it on Spotify, and be forced to compete with a song created by a learning machine? In a time when the average stream only pays out $0.003 and $0.004 per single stream, it doesn’t feel fair that both types of artists get paid the same rates and compete on the same playing field.  

Yet it appears that Spotify has no concerns with AI artists on the platform, as a spokesperson for the platform said they do not “have a policy against artists creating content using autotune or AI tools, as long as the content does not violate our other policies, including our deceptive content policy, which prohibits impersonation.” So, basically, as long as some artist isn’t trying to present themselves as Sabrina Carpenter, it’s fine. They can sound exactly like her because the songs were created by a learning algorithm that was fed her music, but “taking her identity” is where they draw the line. I suppose one could argue that artists are inspired by other artists all the time and that’s what the AI is doing, but it still feels like an intentional oversight by Spotify to have no regulation on AI “artists.”  

It’s also important to note that this is not the first time that anyone has tried to operate under a false persona on the platform. Apparently, Johan Röhr of Stockholm uploaded music under various pseudonyms and sat back, raking in the profits from his 650 personas. Despite the small revenue payout per stream, he managed to bank 15 billion streams, making him a couple million dollars. It wasn’t clear how long that amount took to earn, but it’s still a wild scheme. The difference between Röhr’s method, however, is that he at least created the music himself.  

Perhaps it’s hyperbolic to think that these people making easy music to profit off are really hurting the livelihoods of other artists on the platforms given how little is made on places like Spotify, but it’s really just the first step into a tumble of many creatives suffering because of people turning to free AI tools over the real work of people. Newton-Rex, the CEO of Fairly Trained (a nonprofit that certifies AI companies), believes that many AI music generators are “copyright laundering.” Many people agree with Newton-Rex that AI music generators “exploit the work of real human artists by repurposing their music as training data” (Source). Songs fed into these generators as models and inspiration are being used without the consent of the original artists. It makes you wonder how far copyright laws should reach. If a person can’t take a song and repurpose it as their own without getting permission from the owner, why is it okay for a computer to do so? 

What it also comes down to is the fact that machine learning can only create from things that already exist, meaning no truly original song can be created from it. It will always be a mishmash of existing songs. It can only learn what already exists. This leads back to the question of copyright. Is it infringement for the music generators to take an Olivia Rodrigo song and a Taylor Swift song to create an AI-generated track in that same style? Readers can decide. 

While there can be a lot of useful AI tools to aid the music industry, we have to be careful how far those tools go. We must not let the ease of AI keep us from funding the work from real, human musicians. We cannot let the soul of music die out to improve a music company’s yearly sales. I hope that a post like this doesn’t feel out of date in a couple of years, but at the rate that AI is developing and the rate at which people are substituting AI over human creation, it’s likely that the concerns we face now will be normalized in just a few years. What will that look like for the music industry? Only time will tell.  

But what do you think about the rise of AI music and AI artists? Let me know in the comments!  

Written by Kristen Petronio

Thanks to the following resources for helping me put this blog together. 

https://council.rollingstone.com/blog/the-impacts-and-disruption-of-ai-on-music-industry-stakeholders 

https://www.fastcompany.com/91170296/spotify-ai-music