These songs shouldn’t exist.
Though I’ve dabbled in music for years, I’d never be able to create such extraordinary tracks. From high school orchestra to picking the banjo for my church congregation, and from busking fiddle at farmers’ markets to blues guitar classes at Berkeley’s Freight and Salvage, I’ve tried my hand at various instruments and styles throughout my life. I adore music. Always got a tune in my ear. But I’m no Musician with a capital M.
I’m not a singer or a song-writer. And I’m not a producer. I hold such talented folks in the highest esteem, and I would never want to cramp their style, or threaten their art.
And yet I went and done something bad.
It was so fun though. It was magic. And I fell completely into it, following my own creative spark.
In a matter of days, I created a whole album of AI generated songs.
The notion came to me after I finished my proof-of-concept trailer for The Lost Years. I had used Suno to create the Southern Funk-Guitar track to which Young Yeshua evades the Roman Empire while hunting the secrets of Ancient Egypt, and I had credits left over on my account. (For $10 a month, you can create a whopping 500 songs!) And, I don’t know, I just wanted to try it.
I wanted to hear what it would sound like. I wanted there to be music in the world that spoke to the perilous and potent times we are living through. Songs that faced the metacrisis with full spectrum emotionality, without backing down. An album that was a kind of sonic initiation into this paradigm shifting moment.
So in collaboration with Anthropic’s Claude, I honed my vision for the album and architected its structure following the stages of my own dark night journey: From realization of climate catastrophe and civilizational collapse, through breakdown into the wild edge of sorrow, and rising up to offer my gifts in service to the healing and regeneration of our planet. Then I crafted a custom Gem in Google’s Gemini, a “singer-songwriter” tuned to the frequency of The Great Turning. To create each song, I poured my thoughts and feelings around a particular topic into this Catalyst Bard. This soul-less intelligence pushed back, asked questions, helped me go deeper into my own self-inquiry. Then, when we were ready, Catalyst Bard suggested lyrics and a detailed style prompt for Suno.
Song by song, I iterated to find the right words, the right sound to express the stages of my metamorphosis. I’d challenge Catalyst Bard: Where’s the conflict? Where’s the breakthrough? Make it darker. Make it catchier. Needs more guitar. (I’ll have to be honest, it was fun as to hell to play at being a music producer.)
Sometimes I’d find the song in the first couple of tries. Sometimes it took 30 or 40 iterations, bringing the lyrics and the prompt back to Catalyst Bard for revision and fine-tuning. And then I’d bring them back again.
Generate and iterate. Generate and iterate.
Until, finally, I’d hear it.
Yes, that’s the one!
The instruments are synthesized. The voices are amalgamated from every singer on the internet. But holy fuck does that track slap.
The first time I heard ‘The Flood,’ I melted into tears. ‘Rise Up’ gets my blood pumping for the revolution.
Here was my story, my own personal journey, reflected back to me. And my whole psyche felt the resonance.
Others I’ve shared the album with have had similar reactions. They were deeply moved by the music. Even said it was genre-defining. They were thrilled at the prospect of flooding the airwaves with propaganda for the more beautiful world.
Still others were utterly repulsed by what I had done. And rightfully so! In a couple of days, and with some clever prompting, I had created something that would have taken an honest, human musician months or even years. Where was the artistic integrity in that?
In ‘The Pill,’ my gospel-funk singer belts out: “We’re taking back the soul!” But there’s no soul in that AI generated voice.
Is there…?
The Inquiry
Now I’ve been a struggling writer for over a decade. I could paper the walls of my apartment with all the rejections I’ve received from literary agents and sci-fi/fantasy magazines. The submissions I’ve made that never got a response would light a bon-fire for days. I did sell one story to Analog Science Fiction and Fact, for around $1200. Divide that by the number of hours I’ve been honing my craft, and multiply by the mass of white-collar workers who are about to lose their jobs to AI, and I might just be scraping by on a living wage. (If only they stopped building high-rises in Berkeley, and driving up the rent!)
All joking aside, I know that you don’t make good art without cracking a few eggs. At least not art that’s authentic and sings from the core of your truth. There will always be those who love your work, and those who hate it.
But the strong reactions about AI music, from people I respect and consider friends, prompted a deeper inquiry.
Should I even try to share this music? Or should I just keep it to myself? And how would I bring it forth in integrity, without threatening the real musicians who have poured years into their craft, like I’d done with writing?
I’ve been holding this inquiry for weeks, wrestling with it, and seeking counsel from mentors and friends. I still don’t have an answer.
And yet AI music exists. It’s not going away.
The Instrument
This little thought experiment has helped me make sense of the situation:
Suppose I’d never touched an oboe before in my life. I picked it up, pursed my lips against the reed. Clenched my fingers over the buttons, and—
BLAAACHTT!
I belched out a sloppy, irreverent sound.
But that same instrument in the hands of a virtuoso could emit the most etheric and subtle melodies.
Now imagine an instrument where the pipes and valves are infinitely more complex than the ones that compose an oboe. You could place your fingers on a board of keys and make your best attempt at self-expression. Then the energy of your intention would zip through its infrastructure to output something along the spectrum of slop to high art.
The instrument can’t play itself, though. Even if all those wires and pipes are a million miles long, trace them back to the source and you’ll find a human soul.
Parables of Rupture may have been made in collaboration with AI, but I’m the one doing the belching.
And to the point of integrity: The vast majority of AI generated music out there is sneaky. You might not know that you aren’t listening to a human voice, as artists are not in the habit of disclosing that their work is AI generated. I really think there should be a different category, a choice. But these tools are evolving way faster than we can adapt.
My response is full transparency. This essay details my process. My artist’s profile will name the tools I used.
Listen. Or don’t. It’s up to you.