ChatGPT vs The Creative Soul - Undividing #40

Yes, I use AI. But probably not how you think.

Welcome to Undividing where we are reconnecting a divided world

“If no human could be bothered writing it, I don’t see why any human should be bothered reading it.” - Karl

Hello Everyone,

I feel like I open every Undividing and Emotions Diary with this sentence, but what a joy to be able to. Welcome to all the new folks who’ve joined this week. This idea and community of people undividing the world just keep growing.

To all the new paid subscribers, huge and massive thank yous to you as well. I can’t tell you what it means to me. Someone even gifted a paid sub to someone! Amazing!

So you new 5 paid folks (and everyone who joined this week and is on a 7 day free trial) the Lives every Wednesday are for paid subscribers.

1st and 3rd Wednesdays are for deep dives into the Undividing topic of that week. 2nd and 4th Wednesdays are The Emotions Diary Lives where we do diary entries live and see what the universe has delivered us, share it with each other, and learn more about undividing ourselves on the inside.

• THE WEDNESDAY LIVES HAVE CHANGED TIMES - hopefully more people can join - 6PM Berlin / 5pm UK / 12noon NY / 9am LA. Same link as always.

AI and the creative soul

I attended a live by

Claudia Faith

this week about AI and our work as writers. The conversation there told me how we are really divided as writers and readers on this platform about if we should even be touching AI at all.

Regular readers will know my article about Could AI Finally Bring World Peace? - a moonshot about the global potential of what AI could be in generations from now.

But it’s 2025, we’re writers, and the next article is due. And when I open up the news, my first thought is that is more important than ever for all of us to be writing and putting our throughts out into the world.

The lingering spectre though is AI, if we’re using it, and how.

So I thought I’d write about my journey with AI. It’s become an incredibly valuable tool for me. But I’m very specific about how I use it. The biggest thing for me is never allowing it to replace my soul as a writer. That’s the only thing of value I have to offer. I could never outsource that.

But I think AI has also made me a better, deeper, and richer writer. It’s all in how you use it, in my opinion.

If you know anyone who’d like to get in on this conversation, share this post with them.

OK, let’s get undividing…

---

“Not one of these books has a soul.”

This week’s Undividing was prompted by Claudia’s live, and two other things I saw.

One was a note here on Substack that I wish I had saved! Argh, no searching has turned it up. So below is my writing of the parts I remember.

The note was written by a man whose wife is a book editor. He noticed she had this very concerned look on her face as she sat at her laptop. He asked her, what’s the face all about?

She looked up and said, of the 25 submissions she was looking through, every single one of the books has been written with AI. She pointed out repetitions in language between them all, an emptiness to the messaging, a proliferation of seldom used punctuation that AI loves.

She said something like, “I edit the soul of a book. Not one of these books have one. There’s nothing I can do with these.”

Great book editors are like method actors; they crawl into the book, embody the book’s character, and edit from there. So this isn’t artistic pretentiousness on her part. This is core to how she works.

(post script: a kind reader knew the Note and sent it. I said soul, she said voice:)

---

The second thing was this video with Nick Cave talking about his deep concerns with AI. If you don’t know Nick Cave he is an Australian musician and national treasure. A true original and I thoroughly recommend his free email The Red Hand Files, something I find truly inspiring as a writer and human.

Nick Cave has been very outspoken in his stance against it. But recently changed his mind and wrote beautifully about that in The Red Hand Files entry I refer to here where director Andrew Dominik (Aussie director) created an AI video for a song called “Tupelo” by Nick Cave about Elvis Presley - a beloved song by fans.

With great reservations and objections Nick sat down to watch it (Susie is his wife):

So, I watched Andrew’s film, then watched it again. I showed it to Susie. To our surprise, we found it to be an extraordinarily profound interpretation of the song – a soulful, moving, and entirely original retelling of ‘Tupelo’, rich in mythos and a touching tribute to the great Elvis Presley, as well as to the song itself. The AI-animated photographs of Elvis had an uncanny quality, as if he had been raised from the dead, and the crucifixion-resurrection images at the end were both shocking and deeply affecting. Susie and I were blown away. As I watched Andrew’s surreal little film, I felt my view of AI as an artistic device soften. To some extent, my mind was changed. “It’s a tool, like any other,” said Andrew.

And that there is where the great divide is.

AI is a tool. But it can take over if we let it. And like Nick Cave, I’ve also battled with my relationship with it as a writer. So the rest of this article is how I use it and where I landed.

Because Nick’s example is about a visual, not his songwriting. As our dilemma is not over our pictures we use (mine are always from Unsplash) but the words we commit to the page. And our readers.

ChatGPT is Google without ads

Chat GPT is my LLM of choice.

I hear from other writers that Claude sounds more human, and some people I know swear by Perplexity. But for me that’s irrelevant. I have no desire to have an LLM do my writing for me. So I don’t care how it writes. I care what it can research.

When I start any article in a field I’m not familiar with and attempt to undivide, my first step is to consult with ChatGPT. It lets me discover and explore facets of the space I would never have discovered, and find the voices, thought leaders, and publications in the space that matter. Ones I pull links from.

It’s the best research assistant I’ve ever had.

What I love also about this phase, is that I learn so much more than I would have normally searching on the internet. Because no one is trying to take my data, get me to sign up to something that lets them stalk me for years, or bombard me with ads. I can get tactical about what I read while hacking a path around all the stuff that annoys me no end in the internet landscape.

This period is a masterclass for me as a writer, to formulate my POV. And it’s in this phase that I get to play, learn, and explore.

Beat out the article

I always plan everything before I write it. Have since I started writing at age eleven. “You need a map,” as my high school English teacher, Mrs King, would say. And also, I come from a family of engineers and something about a plan thrills me.

Outlining is what a lot of people call it. “Beating it out” is a screenplay term.

It refers to the “story beats” of a film. I spent seven years as a screenplay writer in film in LA. And a further two years here in Europe in TV. Don’t bother IMDBing me, nothing ever got made. 95% of projects never do.

But the beating out is something I need to do myself first. Because I think of all my writing as a story. I’ll pull together all my favorite parts from the research, construct a flow, then feed it back into ChatGPT for a check.

ChatGPT is a cheerleader. It’s not a critical tool unless you ask it to be. So I order it to get brutal on my beat sheet; what angle have I missed? Is the construction flowing? etc. And I always tell it to be unreasonable, and rip it to pieces.

This is where I find counter arguments that I haven’t considered. Which often makes me re-research and re-beat.

LLMs give me more paints to paint with. With them, I follow my heart as I use, mix and smudge those colours from there.

Or in other words, the part where I start writing.

ChatGPT can never be The Muse

Every word you read by me, my soul wrote. And to prove it, this section wasn’t in the beat sheet.

Full confession: once way back in my early days here, I had ChatGPT write an article for me for Substack. Then I posted it. Then hated it. Then took it down.

I felt I had cheated my readers. And worse, cheated myself. Because I’d sidestepped the magic: in the process of writing an article, the muse comes.

I mean this very seriously. I believe in The Muse - inspirational Goddesses of Literature, Science, and The Arts in ancient Greek mythology. That when you start the process of physically writing, knowledge, thinking, a light of some kind visits you whispering in your ear.

Photo by The Cleveland Museum of Art on Unsplash

It’s kind of ironic that the greatest part of writing as a human didn’t occur to me until I was writing as a human.

This is what I, and I think Nick, mean about the soul and the creative struggle. The visitation of the muse is my personal understanding of how writing works from forty years at the keyboard. That is the soul of writing. You can’t fake it. It’s the place from where artists say a vision appeared, a poem arrived, a song occurred to them. You must simply surrender to it.

When I write something there is always a sense of an unknown to be discovered. And I have no fear in that. An outline is a map, but no map can show you the verdant beauty of the flowers, the mists over the hills that beads on your skin, the music of the forest’s branches and beasts. And then how you convey that on the page so that the reader builds that same feeling and image out of their own storehouses of memories and imaginings. You have to walk it, or in this case, write it.

There is no way on this green earth that an LLM could have written my book, How To Burn A Rainbow, the story of my divorce from my ex-husband. I still read things in there and think to myself, I didn’t write that, the muse did. And it’s why HTBAR keeps touching people from all over the world the way it does.

I have no desire for my experiences to be written in the words of an amalgam of stolen words from other writers. But I do want my writing to be lifted up by the volume of research I could never have performed as one person with a job, a life, and only a certain number of hours to create in a week.

A note on Notes, balls, and spaceships.

Anyone who has ever asked me how Undividing did so well so fast on Substack, I tell them the same three things. This is how I grew it. Which is not everyone’s goal here. Mine is, since I’m to get this message as wide as possible.

• Pick a topic you are relentlessly passionate about.

• Post to a schedule that you never break.

• Post three Notes a day.

That’s it. And I’ve found over time that my Notes general fall into four categories that continually get the most interest; divisions, kindnesses, ageism, tech/unplugging - with an overlay of how this connects or divides us.

And for these topics, ChatGPT is an unending source of new explorations.

My first Creative Director in advertising once said to me, “How many ways are there of looking at a ball?” The answer is of course, infinite. So I use my LLM to Starship Enterprise new ways to see the infinite. Most of what it comes up with is between average and not great. But every time, there is a nugget of genius. A new door it opens. A reframe I’d never considered.

Like the map analogy, this is a signpost. And from here, I dive in and explore.

I also believe that the smallest thing you write must be in your own words. The Muse is my co-writer on every Note. But LLMs are my research department.

They help me unearth more of the infinite. And I write them like I was saying it to a friend.

ChatGPT has turned my advertising life from being a musician to a producer.

As regular readers know, I’m a copywriter by trade. I started nearly 30 years ago in Sydney and my advertising career is the reason I’ve lived and worked all over the world. And like every copywriter I’ve been planning my escape from the industry for most of my time in it; first writing films, then TV, then books, now whatever Undividing will turn into.

But before LLMs I was like a “musician.” I had my way of writing ads that was my style, my thinking, my strategies, my flavour. Ad creatives would talk about each other’s work they way musicians talk about each other’s playing styles or songwriting.

Now though, I write a lot with ChatGPT. It’ll cough up a lot of concepts and headlines that are useless, but in there you’ll always find a gem. As a “music producer” I see this. But I can only recognise it and know what to do with it, because I’ve played the writing keyboard my whole life.

Most producers in the music world are themselves amazing musicians. I’m only a good producer, because I spent decades learning an instrument.

But here is the dilemma; the ad industry is imploding. Freelance is getting scarcer, friends are taking hourly rates over day rates now, but the volume of deliverables being demanded in the time you’re allotted has climbed to the point where you can only produce at that level with an LLM. Ad folks are going through the same thing writers here are. Should I or should I not?

So in my professional life I think of ChatGPT as my team of junior writers. I brief them while I do my own writing, then I see what they came up with that could be polished.

A last thought on LLMs and the creative soul

The AI debate that raged on Substack a few months ago I didn’t step into. It was very binary - you’re either 100% for it or 100% against it. Plus a lot of writers were, I think, very scared to admit that they did use it in some way.

You were either a hack/thief/content factory or a pure artist. It’s more complex than that.

Photo by Jezael Melgoza on Unsplash

I think of it as like when digital photography first arrived and so many of my photographer friends either jumped on it, or swore they never would. Now the photo world is all about how you mix those two as your expression of your art.

I think we’re in a phase like that now. Because like digital photography, LLMs aren’t going away. But we each must find where we’re comfortable holding both.

I hope that this article has helped you undivide your thinking on LLMs as a writer or reader here on Substack. Everyone has to find their own place to sit.

My place is this; I will research the great unknown in ChatGPT. I will have it challenge my work. I will use it to traverse the great unknowns when I’m exploring a new or well worn subject to me.

But the words on the page must be from that magical place the muse speaks from, whispered to me and only me, then placed here for you and only you. And if she drives me off the map, then all the better. That’s where the great stuff usually lies.

I have zero desire to be a participant in writing’s death. But I have a never ending desire to continue to be a better writer. And a writer needs research, themselves, and the muse to be that.

Because if we aren’t writing from this place of wonder, this atmospherical battleground that is creative struggle’s stage, we are not participating in this inexplicable thing called life.

And writing is when I know I am truly alive. And you can’t ChatGPT life.

---

How can you support Undividing?

Like, comment and restack! It all helps the algo push Undividing into people’s feeds.

Have you learned anything in Undividing that’s helped you in your life? Given you a new perspective? Helped you feel connected to the good in all of us? If so, consider taking out a paid subscription. If we can get to a 100 paid subscribers, the algo will supercharge Undividing and get that feeling you’ve had to more people here.

Plus you can join the Undividing Lives every Wednesday for paid subscribers; deep dive discussions into article, and The Emotions Diary live sessions.

You can also Buy Me A Coffee. I post you back a personal 30 sec video for any donation, any size. (Ignore the BMAC amount prompts, you can enter any amount)

You can also support by picking up a copy of my award-winning book How To Burn A Rainbow, the story of how my divorce set me off on a journey - a riches to rags rollercoaster from LA to Berlin where I had to lose it all to find myself.

---