The speed at which artificial intelligence (AI) is improving is astounding. A month ago, we were talking what-if scenarios. Today we’re talking about reality.
A few days ago, Ghostwriter released their song Heart on My Sleeve. It’s a song that used the voices and/or music of Drake and The Weeknd as training for an AI. The details of how Ghostwriter created the song is currently unknown. But Universal Music Group wasted little time with takedown notices sent to the major platforms that posted the song, including TikTok, Spotify, Apple Music, and YouTube.
But even if that hadn’t happened, we could logically argue that data used for AI training and the musical outcome is analogous to samples used for songs. Examples include Janet Jackson’s Rhythm Nation sampling of Sly and the Family Stone’s Thank You and MC Hammer’s U Can’t Touch This sampling of Rick Jame’s Super Freak, to name two. Before that, Jimi Hendrix rearranged Bob Dylan’s All Along the Watchtower. And if you go back to the classical music era, many of the greats used traditional folk and gypsy music within their compositions.
Music has always been one of the art forms at the forefront of change. In the case of AI, photography has already been affected. And soon, AI will be coming to video/movies and other art forms.
Artificial Intelligence and Music Artists
Where does that leave music artists? There are going to be five groups who do well.
Note: Much of this depends on how fast laws can catch up with AI technology. Having said that, I’m not a lawyer, and nothing I’m writing should be construed as legal advice. If you’re interested in a lawyer’s current viewpoint, check out this interview with attorney Chris Mamen.
1. Artists Who Write Their Own Music
As an artist, artificial intelligence is intriguing to me from the perspective of when I write and record a new song based on my own ideas, I can then train an AI to create a dozen or more iterations of my song to find out which version is the most interesting musically and which version is the most popular.
In business terms, this is called A/B testing, and it’s something that’s been used in the music industry since at least Berry Gordy’s Motown Records. For Gordy’s A/B testing, he’d assemble groups of everyday people to listen to various versions of songs and refine them based on their feedback.
Also, releasing different versions of songs with guest artists has been around since the 1980s and was used for artistic and business reasons–specifically to bring more notoriety to a song or artist who wouldn’t usually have access to a particular audience.
Like Dylan adapting the Hendrix version of Watchtower for his live show, I’d do the same with whichever version of my original song was the most popular. And like Gordy, this is a pragmatic view on creating a good song.
It would also benefit people who need music for movies, videos, or business (presentations, websites, etc.).
“I really like that song, but I’d like a slower version with a reggae feel. My budget is $250, and I need it tomorrow.”
Currently, that budget and timing wouldn’t work, so the person making the request would likely end up where most people do: a website that sells or gives away pre-recorded (stock) music.
But with artificial intelligence, instead of settling for something from stock music, they could get exactly what they wanted within their budget and time constraints.
2. Artists who have a Unique Sound, particularly Vocals
This is already happening with artists like Grimes. But the precedence was set by Trent Reznor, who historically has encouraged people to make their own music based on his.
3. Historically Renown Artists
Want the next Beatles album from the Sgt. Pepper’s era? There are already Lady Gaga and Oasis songs available created, at least in part, with AI.
But before you start, talk with your attorney. This is where many current and future lawsuits and legal issues will need to be resolved because a lot of money is at stake.
Video of AISIS – The Lost Tapes / Vol.1 (In Style of Oasis / Liam Gallagher – AI Mixtape/Album
4. Everyday People
We’ve all had an idea for a song, but most people aren’t trained vocalists or musicians. Soon, anyone will be able to create a song with the help of AI.
A simpler version is the option on current streaming platforms to choose playlists or songs based on a mood. In the not–distant future, an AI will create a custom set of new songs based on your mood.
5. Human Made Music
In 2023, vinyl records outsold CDs (compact disks). This feat has not happened in a long time.
While most people prefer the simplicity of streaming and happily trade recording quality for convenience, the vinyl niche is worth over US $1 billion.
The same thing will happen in the future with human-made music. AIs will likely make the vast majority of music, possibly with some level of human collaboration. But there will be a market for human-made music. This could go as far as having a standard certification process to ensure that “no AI was used during the creation process of this music.”
Artificial Intelligence and All Artists
While the above may be scary for music artists and photographers, particularly with the speed at which AI is evolving, it’s really going to hit home when movies are made without actors and traditional production.
AI and 3d animation rendering isn’t quite there yet. But it will be soon, particularly with computer graphics hardware companies like NVIDIA and AMD entering the AI space.
There will be niches for certified human-made movies. And similar to music vocalists, actors with unique looks or attributes will be in demand.
But what if you are in the mood for a new Romantic Comedy set in 1920 Iceland? The convergence of AI and 3d animation will be the future. And that change will also redefine the redefined Hollywood Business Model.
Photography credits |
Tom Libertiny (Cover)