Jukebox: The AI of Music

We’re all well aware of the mind-blowing ways AI technology is taking over the universe — smart assistants, self-driving cars. It’s even become a life-saver (no pun intended) for surgeons who can now reach places inside of our bodies that their hands couldn’t get to before. So all good. 

What’s next though? Will psychologists be replaced by robots who can guide us through our emotional conflicts? And would they be capable of passing a box of Kleenex?  

Creators have been using computers, samples and synthesized effects for some time now. But we make the decision as to whether or not to implement them. We’re doing the programming and choosing the sounds. We’re in control. What happens if someday we’re completely cut out of the music-making process? 

I watched an illuminating video (included again at the end of the read) about what Artificial Intelligence Music Sounds Like and it flipped me out. (Ironically it was narrated by Harrison Renshaw who literally sounds like a robot. But he’s not. (I checked.) 

Harrison introduces us to “Jukebox,” a system that intakes the genre and mood of a given song, the artist and snippets of lyrics, and then is given the task of ‘making’ a song that sounds like that artist could have written and performed it themself. Jukebox can mimic voices, instruments, scores, styles, symphonies, you name it. And it’s super vibey.

Uh oh. 

The technology simulated what Kanye West would sound like if he sang Eminem’s “Lose Yourself” and I have to tell you…it sounds  uncannily like Kanye West singing “Lose Yourself”. The quality of the recording is ruff, but these are early days. 

Check out how close they come to Frank Sinatra, Elvis Presley and Katy Perry. The lyrics are gibberish but I have no doubt they’ll get more coherent as the bugs are flushed out. 

What’s really freaky is that the Katy Perry ‘song-clone’ sounds like a smash! I’m not kidding. Any crafty topliner could find words to replace the gibberish. It even manages to capture the Katy-clone’s yearning and lusting. 

I have questions. What are the intellectual property implications here? Who owns the rights to the song-clone? I mean, I’m sure Jukebox would, but what about the technician who chose the specific language or keywords to input in order to get the desired sample? Can’t he/she/they claim some credit? 

Can Katy Perry sue if the ‘performer’ sounds too much like her as Bette Midler did for a vocal impersonation on a 1986 TV spot? Composers know that the line is thin. You want a sound-alike but you run a legal risk if it sounds too much alike!

OpenAI, an independent research organization founded by Elon Musk, got billions of dollars to develop Jukebox because “they want to use computers to make the world a better place.” (That, my friends, is subjective.) 

This isn’t artificial intelligence. It’s artificial art. And that’s an oxymoron. Art is the expression of human creative skill and imagination; works created by artists. Do we want the portraits on our walls to have been conceived by a computer? Or poetry about the sweetness of life spat out by a machine that never tasted it? Swan Lake performed by mechanical dancers? Or…does any of that matter if in fact it makes us happy? If we can dance to it?

There’s more. As if songwriters aren’t struggling enough with dwindling digital royalty rates and competition for song placements, will we now have to compete with thousands of compositions composed by non-humans as well? I feel fortunate that I never had to think about this stuff as a young songwriter. Diane Warren was my biggest fear! I’m more grateful every day that I came up in an innocent time when a Jukebox was a music-generating device in a diner into which we dropped a coin and pressed B5 or C8 to hear out favorite tune.

Harrison imagines that in the future AI music will be in a genre of its own. He goes so far as to suggest that maybe robots will be making albums (sexy!) from scratch that sound just as good as the ones we’re listening to now. Yeah but, we’ll know it’s from a robot. It’s like having sex with a blow-up doll — she doesn’t really care. And how dull a concert would be without moving, shredding and sweating. 

In his new book “Keep Sharp: Build a Better Brain at Any Age” (one of my favorite Covid reads) Dr. Sanjay Gupta writes, “No matter how sophisticated artificial intelligence becomes, there will always be some things the human brain can do that no computer can.” 

Jeez, I certainly hope so.  😳

Thanks for reading, my friends. Listen to my album 2.0 etc…If you'd like to receive my blog via email, please click here. Follow me on Twitter and Insta. Visit my Serial Songwriter Facebook Page. Get a signed CD or a copy of “Confessions of a Serial Songwriter.” ☮️