I have used a version of “artificial intelligence”—AI—in audio production for a few years now. That is not to say I have created actual music, parts, or performances using it. The seeds of AI in my productions were planted by mastering plug-ins that listen to a recorded track and suggest settings for the best sound. It gives me a good starting place for finalizing my mixes.
Much of what it does is simply a time saver, perhaps no different than manually creating a template, except the template is created automatically by an algorithm listening to the mix. I can save its suggested starting points or have it listen again and give me another.
As a professional, I can’t see myself using AI music production tools for composition or songwriting when I could play the part myself or bring somebody in to play it. But I have no issue with anybody using the tools available to them in the second quarter of the 21st century. What AI can do with music and voice is in its infancy, and watching its evolution has been fascinating.
My preference is that people interested in making music learn to actually make music. But I have seen a reverse psychology of sorts in young people learning to play, sparked by technologies often ridiculed or discounted by musicians and music teachers.
I think of The Beatles: Rock Band video game. My youngest son learned to play the Ringo parts at expert level. He did the same with the guitar controller. The game helped him develop a love for The Beatles—winning!—but it did something else. Playing drums in expert mode essentially meant playing the original drum parts. Play them wrong, or let your timing slip, and your score suffers.
My son was scoring perfectly on expert levels. Over a year, he essentially became a human metronome with perfect timing. Maybe he already had it inside him and the game brought it out, or maybe the game trained him. I don’t pretend to know which.
When he was 13, he asked for a Hofner Beatle Bass for Christmas. I bought him one. He dove in headfirst using tablature, as did several middle-school friends with whom he is still in a band today.
They learned “in reverse,” as I call it. They learned from tablature on a screen, with a cursor moving along with the music. They ignored the written notes and watched numbers on a fretboard. They knew nothing about theory, what they were playing, or why it worked with the other parts. What they did have was an accelerated learning path using modern technology, sparked by playing plastic guitars in a video game.
Nothing sparks interest in playing like hearing yourself succeed.
I came home from the winter NAMM show six or seven weeks after my son got his bass. Walking upstairs, I heard the distinct bassline from “Lady Madonna.” I opened the studio door and found him leaning forward, staring at the screen, playing the bass part while watching the cursor move and never looking at his hands.
I asked what he was doing.
“Dad, it shows you where to put your fingers.”
He was, in essence, sight-reading tablature. I challenged him to play a song he had never heard by reading the tablature. He did.
A few years later, I tried explaining the Nashville Number System to him and his friends, hoping to sneak in some theory. One said, “I understand what you mean by ‘the one,’ but which string?”
Clearly, I needed to start at the beginning—a very fine place to start. I started explaining Do-Re-Mi so they would understand the first, fourth, and fifth notes of a scale as the one, four, and five.
They had never seen The Sound of Music. Sigh. I am old.
I explained that knowing the one, four, and five meant they could play songs using those three chords in any key. They had no idea what I was talking about.
Yet here we are 15 years later. Those kids are still playing together, and their affinity for making music eventually drove a natural curiosity about chords, scales, and theory.
That is why I am not particularly concerned about people making music—even entire songs—using AI. Being dishonest about how you made it is another discussion. But if creating music with AI provides gratification, success, and encouragement, history suggests something interesting may happen. Students genuinely interested in becoming musicians may eventually become dissatisfied having a machine do what they want to do themselves.
AI for music, videos, or lyrics is like spices in the pantry—something that might help you learn to cook. Some people never graduate beyond warming frozen dinners in a microwave. Others decide they want to learn how that meal is actually made and disappear down the culinary rabbit hole. That is very human. So let the kids use the technology. See if there is a spark to ignite.
We are not talking about the students already choosing instrumental or vocal classes. We are talking about those who probably never would. If a student in Music Appreciation or General Music learns to prompt musical ideas, even if a computer initially generates them, those who feel something may eventually want to create organically.
And those who don’t? They may never have pursued musicianship anyway. But they can still learn song and lyric structure by observing the relationship between their ideas, prompts, and the AI’s results.
Ignite the spark using the tools they are growing up with, not necessarily the tools we grew up with: printed pages and instruments.
The AI toothpaste is out of the tube. Make peace with it and flip the script. Take that perceived “easy way out” that supposedly isn’t “real music” and use it to teach what music is made of. Teach students to write prompts requiring an intro, verse, chorus, bridge, ending, and other musical elements.
From that angle, this whole AI music thing might look a little more encouraging in your classroom.
Mike Lawson is a musician, and publisher of MMR and SBO+ Magazines
