From Conductor to Curator: The Changing Role of the Musician in the Age of AI

I’ve been playing guitar off and on for about ten years now, with a few classes thrown in
there. I’m still not good enough to feel I could cut an album. More crucially I have
absolutely no ability with my voice. These limitations aside I just released my first album,
Repent. Eight tracks that (I think) sound professional, polished, and intentional. I made it
using AI tools, and the process gave me a lot of thoughts on the current state of LLM’s and
their impact on our future.

"Repent" By George & The Tumbleweeds

 The Process
Here's how a song comes together for my project, George & The Tumbleweeds: I usually
start with ChatGPT, either feeding it a concept or a lyrical fragment I want built into
something larger. For Repent, the title track grew from a single line I had in my head — "I'd
repent to God every day just to be your lover." I gave that to the AI and shaped what came
back. Other songs like "Favorite Bad Habit" and "Me & My Tumbleweeds" started as pure concepts that I guided and edited into final lyrics. A few songs were written wholly old-fashioned, including Rewrite History.

Then comes the music itself. I take the lyrics to Songer, select genres and stylistic tags, and
generate. The program gives me two samples, sometimes the right version appears in one
or two attempts. Other times — like with "Rewrite History,"— it takes two months and
dozens of generations before the output is something I can accept as my final song.
The job isn't executing the vision with trained hands anymore. It's holding the vision clearly
enough to recognize it when it finally appears.

Historical Precedent
This pattern isn't new. The tools change, but the shift follows the same shape every time.

When Gutenberg's press arrived, the production of books moved from scriptoriums to
workshops. The skill that mattered was no longer penmanship — it was editorial judgment.
Deciding what to print became more important than the physical act of printing it. 

When photography emerged, it didn't just threaten painters. It redefined what painters were for. The ability to accurately render a scene became less valuable overnight. Interpretation, abstraction, and vision became the differentiators.

When recorded music became possible, the live performance was no longer the only
product. Suddenly the choices made in the studio — arrangement, production, mixing —
became their own art form. The role expanded upstream.

Each technological shift moves the human contribution earlier in the process. The
execution becomes automated or democratized. The conception and curation remain
human — at least for now.

The Translation Layer
There's a practical reality that anyone working with generative AI learns quickly: AI-
generated content often makes better input for other AI systems.

When I feed human-written lyrics into Songer, the output frequently fights the words. The phrasing sits awkwardly in the melody. The structure feels forced. But AI-generated lyrics flow naturally into the music — the rhythm aligns, the syllables land where they should.

This isn't unique to music. Using AI to write prompts for image generators like Adobe Firefly produces better results than human-written prompts. The systems speak a common language. Human input often needs translation.

This raises a question worth sitting with: as AI tools become more integrated into creative workflows, how much of the process becomes machines talking to machines, with humans serving primarily as editors and approvers?

Addressing the Obvious Criticism
"That's not real music." "You're not a real musician." These responses are predictable, and they're worth engaging with directly.

The question is what exactly we value in music. If we value the discipline and mastery required to play an instrument — the calloused fingers, the years of practice — then AI-assisted music fails that test. That's a coherent position.

But if we value the final work and its effect on the listener, the origin matters less. A song that captures the specific ache of heartbreak doesn't become less affecting based on who played the notes. And if we value human expression and vision — the album Repent came from real experiences. Real late nights. A real complicated relationship with faith that became "Never Learned to Pray Right." The AI didn't feel those things. I did. The AI just helped me build a container for them. 

The line between tool and collaborator has blurred. Whether that's a problem depends on what you think music is fundamentally for. 

The Industry Implications
The barrier to entry for producing professional-sounding music has collapsed. That's going
to flood streaming platforms with content — most forgettable, some genuinely good, and
no easy way to tell the difference before listening.

Production skill is no longer scarce. What becomes valuable instead is vision, taste, and
the curatorial judgment to know what's worth making. The ability to hold a concept clearly
enough that you recognize it when the machine finally gets it right.

Live performance remains human. The presence of a person on stage, connecting with an audience in real time, isn't something AI replicates. Musicians who master both the traditional craft and the new tools will likely create work that neither approach could produce alone.

But for those of us who can't play piano, can't sing on key, and never would have made an album under the old rules — the door is open now. The role has changed, but it's still a role.

A Note on This Article
There's a fitting irony here: I wrote this piece the same way I make music. I gave Claude AI a concept and some raw material. I answered questions to refine the direction. The AI generated a draft. I edited, pushed back, requested revisions, and shaped it until it matched what I was trying to say. The process for writing this article is nearly identical to the process for producing a track on Repent. 

If that bothers you, consider whether the argument above changes based on who typed the words. The ideas are mine. The curation was mine. The AI helped me build something I couldn't have built alone — or at least, couldn't have built as well. 

That's the job now. From conductor to curator.

By George Tillman

George Tillman is the creative force behind George & The Tumbleweeds. His debut album "Repent" is available now. Check it out Here

From the Publisher:
Eight Buffalo Media Group is pleased to support fellow creators in sharing, publishing, and promoting their work. Hosting this announcement and sharing this comic strip is just one example of our commitment to the larger creative community, and we look forward to championing even more voices in the future.