Will AI replace musicians?
Humanity is already biting its nails over artificial intelligence spreading influence in every area of life. But not in a thousand years could we predict that AI will be able to make music better than musicians themselves. The media is taking it as far as claiming that AI is better at making Drake songs than Drake himself. And you probably know that Drake isn’t the one to take that comment easily.
Now, before adding to this narrative, let’s clarify one thing.
AI isn’t better at making music than humans. We’re just fascinated by how well it can mimic the work of people that we’re already familiar with. Just like in any other industry.
In this article, we’ll go over how AI-generated music works in a way to deliver songs as if they’re made by established musicians. Moreover, we’ll look back on some of the recent events on the topic, such as the album “AIsis”, referring to Oasis, or the viral Drake song created by one TikToker, as well as Grimes’ philosophy of why AI-generated music isn’t such a bad thing after all.
Buckle up!
Oasis AI-generated album “AIsis”
Artificial intelligence has already been used in the music industry for a while by some big names, such as James Blake. The thing is, projects like this were mainly used as a form of so-called adaptive music. Think of it as personalized soundscapes, where the sounds adapt to the environment, time of the day, heart rate, location, weather, and more.
However, the first time AI-generated music became viral happened only recently, with what is presented as the new Oasis album.
As you probably know, Oasis was a rock band from Manchester, mostly known for their song “Wonderwall”. The band fell apart in 2009, as a result of a conflict between two brothers, Liam and Noel Gallagher, who later became known for their eccentric and quarrelsome public behavior. Either way, their fans worldwide were disappointed the members couldn’t overcome their disputes over the sake of music.
But, AI made that possible, in a way.
So, what exactly happened?
Mid-April 2023, an AI Album was released on YouTube, called “AISIS – The Lost Tapes”, and presented as something Oasis would make if they were still playing. As expected, fans lost their heads over it.
But what wasn’t so expected was the fact that everyone was quite amazed by it. Even those who pressed play only to hate it afterwards, still didn’t have enough material to say anything bad.
If you scroll down the comment section, you’ll see honest impressions of fans.
Now, let’s answer the most intriguing question: How was this album made, technically speaking?
You’d be surprised to learn that humans made it, actually. And not just any humans – the band Breezer wrote, played and recorded all the songs, including the vocals. The vocals were then pulled through AI, which has been fed with data on Liam’s voice, replacing the original singer’s voice and cadence.
In other words, you’d need to train AI with a bunch of a capella recordings.
AI is still very much controlled by the user. You need to feed it exactly what it needs to replicate. I don’t think it’s at the point where AI could write a song.
Producer Bobby Geraghty. Source: The Guardian
This is just to show how artificial intelligence could be used as a powerful tool to enhance music production, if done properly and with purpose.
Viral Drake song made by a TikToker
Now that you’re familiar with the basic concept of it, you can go on and explore the internet, only to realize that people have been doing this for years. Just type in AI-generated music, and you’ll come across Kanye West singing Lana Del Rey, Gotye, and many others.
It’s true that, as with most AI-related things, this field isn’t as regulated yet. That’s why there are so many examples online, but one artist in particular didn’t want to let things slide.
If we learned anything from Drake’s career, it’s that you don’t mess with him and simply get away with it. Especially if you attract as many as 15M views on TikTok.
But in this paragraph, we’ll talk some more about how can individuals who created music using AI actually be sued for recreating somebody else’s voice, based on the example of the song “Heart on My Sleeve”, purporting to be Drake and The Weeknd.
Namely, this track was originally published on TikTok by the user Ghostwriter977 who also shared it on streaming services under the same name. That was probably his mistake number one, since music services such as Spotify are much more regulated compared to YouTube or TikTok and are therefore much faster to react when something like this does slide, which rarely happens.
Not only that, but the song went viral over the weekend, reaching up to 600,000 streams on Spotify, 275,000 views on YouTube, and the mentioned 15M views on TikTok. This astonishing success only made the latter reaction faster and more rigid.
The Universal Music Group (UMG) managed to pull the songs from streaming services, condemning it for infringing content created with generative AI, but the damage has already been done.
The training of generative AI using our artists’ music (which represents both a breach of our agreements and a violation of copyright law) as well as the availability of infringing content created with generative AI on DSPs [digital service providers], begs the question as to which side of history all stakeholders in the music ecosystem want to be on: the side of artists, fans and human creative expression, or on the side of deep fakes, fraud and denying artists their due compensation.
Source: UMG Spokesman statement for Billboard
The core issue here is that you cannot train AI on copyrighted music without obtaining the required consent. So, anyone who’s been playing around with this potential of AI may be subject to legal consequences.
On the other hand, some artists did agree for their voices to be used this way.
Why is Grimes supporting AI in music?
Now, the real plot twist lies here: copyright law refers to the idea of making a copy, which AI-generated music is not. The only legal reason the song “Heart on My Sleeve” was pulled from streaming services was that it contained an unlicensed sample of the producer Metro Boomin’ saying his name, which was protected by copyright, according to the BBC.
In other words, regulation will definitely have to keep up with all the grey zones that AI is creating nowadays, and it needs to act fast.
Contrary to most opinions, Grimes was the first to publicly support the use of AI in creating music. As a matter of fact, she stated she would share 50% of the royalties with the creator, the same way she would with any other collaborator.
Not only that, but she claimed it’s interesting to be fused with the machine and the idea of open-sourced art.
This doesn’t come as a surprise, considering that the musician was previously married to Elon Musk, and has vastly explored the idea of using AI in music, as well as the ethical boundaries surrounding the human-machine relationship.
So, to end on a brighter note, perhaps AI should be considered as just another tool that should help musicians and producers to upgrade their skills and creative process. It also has the potential, like the example of AIsis showed, to bring back some artists to life, which definitely has a sentiment to it that the global audience would enjoy, and therefore could unlock a new revenue stream.
But keep in mind that earning profit out of somebody else’s name is never okay, whether it includes AI or not. So, if you’re interested in creating music using AI, make sure to ask the artist for consent.
