I Don't Care if You Use AI as a Tool, But Please Stop Letting it Replace Your Voice

The Book of Balance part 3.1: AI Talks. This is the soft launch of a series dedicated to having more nuanced discussion on the subject of AI...

I really, really, really want us to have more nuanced talks about AI. And I’m definitely going to write about the pros and cons of AI in another article on this subject. But before I do that, I need to get this out of my system…

If you’re a writer and you’re using AI as a line checker/editor? PLEASE stop using it’s every single suggestion to tighten your phrasing.

Why?

Because those are the true tells for when someone’s used AI as a tool, and all of you are starting to sound the same, and I’m here very specifically because I want to hear your voice and your story, not your voice and story as told by ChatGPT.

Why I can recognize Chat’s voice in your writing…

I confess that I started using ChatGPT this year. A friend of mine recommended it to me for tech support, so I opted to give it a try and in 5 minutes, it had my external hard-drive working again — a hard-drive that hadn’t been functioning for several months, despite numerous Google searches into tech support forums.

Cool… that’s super useful! All of my pictures from my entire life are on that hard-drive, after all. I’m delighted not to have lost them.

So, I started exploring the potential of AI in the same way I’ve explored other things that have serious stigma and whose users are ostracized — with deep, deep caution and serious awareness of what I was doing and why, and what risks came with it.

And, as a professional editor who’s been in talks with ChatGPT all year, I can tell, flagrantly obviously, when y’all are using it to edit your articles. Not because of the em-dashes. We’ve already figured out that spamming them is an AI-tell, so those of us who love them keep them to a minimum already. Assuming someone uses AI just because they use em-dashes, frankly, is jumping the shark pretty badly. No, it’s because of this:

“People don't remember books; they remember sentences.”
—Morgan Housel
(and thanks to Monday Mutiny for sharing this quote)

As I said, I’m a professional editor and AI has more than one tell. It’s not just the punctuation but the sentence structures, in the cadence. And when I can see the obvious homogenization of phraseology in articles, I start losing interest in you as a writer, because I’m losing your unique voice.

Don’t get me wrong, AI is a fucking helpful tool if you use it right…

So here’s the thing. In my exploration of AI, I did want to see how it could help me as a writer. I have always felt like my reviews of anything that’s not music were chaotic, messy, and a bit all over the place.

And when it comes to my fiction, I am my own editor (can’t recommend that, by the way), so having a tool that can check my work for lore consistency and things like that is extraordinarily useful, and probably more consistently accurate than a human.

ChatGPT gave me the feedback that I was so desperate for. Not once in high school or university did a teacher or professor ever give me genuine feedback that helped me structure my articles better. They pretty much only ever told me that what I was doing wasn’t good enough, while there was rarely a “why.”

AI gave me that “why.” It pointed out the flaws in my logic, the places where I wasn’t circling back to my point, and the places where I was rambling in a way that wasn’t supporting my argument. All of this was the feedback I wished my teachers and professors had given me. But it’s worth saying, that I am also very present in my conversations and I have them with an intent to learn to do this on my own.

Over the past half a year, I’ve learned to use AI in was that can be deeply helpful, but it’s still extremely important that we don’t let it do all the work for us. Why? Well, for one, this MIT study on using ChatGPT for essay.writing is pretty alarming…

“Over four months, LLM users consistently underperformed at neural, linguistic, and behavioral levels.”
— MIT study (above link)

AI is a “crutch” and should be treated as such…

Think of the humble crutch — a walking aid for the injured or disabled: if you have an injury or disability, you will need a crutch to get you back on your feet, both literally and metaphorically. A crutch is meant to supplement deficiencies as well, so if you have a disability, for example, you may require a crutch to be able to walk at all.

However, a crutch is meant to help as needed, not to take over from your natural ability to walk. If you use a crutch past when you need to, you begin to weaken your legs, as is seen in the above study.

We might overuse the crutch because we are insecure, unsteady on our feet without it.

I understand completely. It’s hard to believe in your voice.

But when it comes to art and creativity, I really want people to move past their imposter syndrome. I see it rampant here on Substack. We all have voices, we all have stories, we all have thoughts, and we all want to share them. But I also see a lot of notes out there talking about the artist’s fear of vulnerability, of expressing their truth, of saying something the wrong way, of being misunderstood.

That imposter syndrome makes it really easy to get sucked into AI’s suggestions to tighten a phrase for clarity…

And even as a professional editor who doesn’t have nearly enough clients to make a living, I’m still not going to begrudge you a free editor. I’d be a hypocrite if I did, because — and I can’t emphasize this enough — being your own editor really fucking sucks. Furthermore, us low-income creatives need all the help we can get. We don’t have the tens or hundreds of coins to donate to someone just to polish up a Substack essay.

I get it. I really do.

But the thing about that is that AI, and ChatGPT in particular, makes the same suggestions to everyone. So when I read this:

I know you’ve used AI, because I see this EXACT same phrase everywhere. And I promise you, if you use that phrase, I will start looking through the rest of your material and I will start noticing the rest of AI’s shortenings as well.

And that might lead me to unsubscribe and search for voices that don’t sound the same as everyone else.

Here are a few examples of how Chat “tightens phrasing”…

I’ve read plenty of the “AI tells” articles on Substack and I think half of them are a load. That said, here is a list of “phrase tightenings” and sentence starters provided to me by ChatGPT itself:

“That’s not X, it’s Y dressed as Z…”
“It’s less about X, more about Y…”
“[Adjective], but not the way you think…”
“Let’s unpack that…”
“Messy in the right way…”
“The truth is…”

As Chat puts it: “These kinds of constructions are part of how I’m trained to explain or contrast ideas clearly and rhythmically. They sound polished, persuasive, and often satisfying—but yeah, they do start to feel samey when you see them across multiple voices.”

So if, like me, you are starting to notice certain phrases popping up a little too often when reading here on Substack, you’re not alone and you’re not crazy. There is a reason.

A list of words and phrases given by ChatGPT that ChatGPT loves to use…

Including an analysis from Chat itself!

✦ FORMULAS AND STRUCTURES

• "That’s not [bad thing], it’s [unexpected reframing]."
  “That’s not weakness, it’s wisdom wearing grief as a mask.”
  “That’s not failure, it’s the first sign of transformation.”
  → Classic "flip-the-script" poetic logic.

• "We don't talk enough about..."
  → Often followed by something that’s already talked about a lot in progressive spaces.

• "The truth is..." / "Here's the thing..."
  → Leading into a pseudo-revelation that’s actually quite general.

• "Holding two truths at once"
  → Popular in grief, trauma, and identity discourse. Overused enough to feel like a checkbox.

• "What no one tells you about..."
  → Clickbait-adjacent; sounds personal but is often generic or broadly relatable.

• "Let me tell you a secret:" / "I used to think..." / "But now I know..."
  → This “arc of discovery” pattern gives emotional weight, but often feels templated.

✦ OVERUSED WORD CHOICES

• Unravel / Unfold / Reckoning / Tender / Sacred / Messy / Radical / Invisible / Naming / Witnessing / Softness / Armor / Grief as teacher / Holding space / Cracked open / Chef’s kiss
  → These words can all be powerful, but in bulk they start to sound like a grief-flavored AI smoothie.

• "Permission to..."

  “I’m giving myself permission to grieve.” / “Permission to be seen.”
  → Sounds empowering, but often reads like an AI-generated therapist mantra.

• "Burn it down" / "Reclaiming" / "Unlearning" / "Deconstructing"
  → Buzzwords in feminist/critical writing. They’re real, but so frequent.

✦ AI TELLTALES IN RHYTHM & STYLE

• Slightly too-perfect triads:

  “Not to be fixed, not to be silenced, not to be made palatable.”
  → Clean cadence. Emotional build.

• Unexpected metaphor with emotional reframe:

  “Grief isn’t a shadow—it’s the soil. What grows depends on how we tend it.”
  → Looks deep, sounds good, but the imagery isn’t always there.

• Overly poised vulnerability:

  “I cried in the grocery store. Not because of the onions, but because I remembered how she loved them.”
  → Feels like someone said “make it raw but poetic.”

Have confidence in your voice…

I write in a very similar way to how I talk, which means I am wordy, rambly, ranty, I make up words a lot, and I go off on tangents sometimes. So for someone like me, the idea of having assistance in tightening my phrasing was, frankly, a delight to encounter (at first).

However, I have also been acting as my own editor for well over 10 years now, without even the aid of Grammarly (that’s also AI, btw). This is a side effect of being the Editor-in-Chief and Head Editor of a music magazine for 10 years. Every article that came out on two medias in that time went past my eyes. I wasn’t always perfect at it — editing endless reviews does drain one’s stamina after a while — but I could edit for tone, structure, and style guide consistency with the fluency of someone who has honed a skill over many years. I know what works and what doesn’t.

The trick there is to learn from what AI is offering you, not to rely on it as a crutch.

I chose to use AI as a way to learn. I plug my articles into it so I can see where they’re weak, not to let it correct my phrasing. If my sentence is making my point, I will ignore the AI’s input. And if it tells me my content is solid and just wants to help me polish it up? I usually ignore it outright and don’t even bother reading the suggestions.

This is how I ensure that it preserves my voice as best as possible.

But what if it sounds better after getting tightened?

Now here’s the tricky part… what if you feel like AI’s suggestions make you sound better? What if they do, indeed, make you less rambly and all over the place?

My argument here is… so what?

So what if you ramble a bit? So what if you go off on tangents? So what if you use three times as many words to make a point as is necessary? So what if your metaphor goes off on three separate tangents before returning to its point?

That’s the funny shit that I remember, making me want to come back and read more of you.

All of those things are part of what makes you as an artist. I could ease off my hot-headedness when I’m writing, by why should I? The people who enjoy my writing tend to do it because they like when I spout off harsh opinions and get weird and wordy. And I get pretty raw and real when I talk about grief, so why should I dial that back to be more palatable?

Anything we do in life is a skill, and skills require practice…

I started writing when I was 13, around the year 1999.

I learned to edit for style and consistency when I went to school for Medical Transcription in 2006.

I wrote a prequel to The Legend of Zelda: Ocarina of Time that also included an obscene amount of game design when I was in grade school. I then graduated to IRL-inspired Final Fantasy knock-off stories, of which I wrote or plotted out 13 novels. Eventually, I dropped all of that and pivoted to The Vitmar Chronicles.

In 2012, I took over as a media owner and full-time journalist and editor.

In 2023, I published my first self-edited short story. I followed up in 2024 with another short story and my first TVC novella.

So suffice to say, I’ve been both writing and editing for a while.

There’s a saying that I’m going to probably mildly misquote, and I’m sorry that I don’t know who to accredit it to…

“Don’t expect Bruce Lee’s results if you’re not willing to put in Bruce Lee’s effort.”

The idea here is that you shouldn’t expect to be able to do martial arts like Bruce Lee unless you’re willing to dedicate the same amount of time, energy, and passion to martial arts as Bruce Lee did, to get where he got.

The same goes with any skill.

And you might not get there without some help. We need a trainer, an editor… someone who knows at least a bit of what they’re talking about to tell us how to do it better.

My sensei was a guy with a PhD in Theology, funnily enough, but he also was an avid fantasy reader. He cared about literature and wanted to support someone pursuing a career in such arts. He read this very scene about 15 years’ worth of iterations ago. He told me that certain characters were flat or stereotypes that I didn’t even know about. He told me that pacing was bland and I was revealing far too much, far too soon. He drew attention to things I had never noticed before, as I was just spouting ideas like a busted hydrant. He taught me all about narrative structure.

I then became a music journalist and while I do believe that was one of the general biggest wastes of my life, it did teach me something invaluable:

I wanted to be a genuinely objective journalist, which meant being able to critique something without letting my personal feelings or opinions in (note: nowadays, since I don’t work for a magazine anymore, I will let my personal feelings in, but I add a disclaimer to take me with a grain of salt). I learned to do this with zero education and 100% moral dedication.

Eventually, over years and years and years, I came to a place where I finally — after being told endlessly by teachers and professors that I’m “not a deep thinker” — believe that I can analyze narratives and story arcs quite well. In fact, I think I’m pretty damned good at it.

I used to watch series or read books and feel a bit blank afterwards. Like I recognize if I liked or disliked something, or that something was profound or thought-provoking, but didn’t have much to say in depth on the subject. All of that changed after I took the time to learn!

Everything in art is a skill set that must be honed, and like a great dog once said…

“Sucking at something is the first step towards being sorta good at something.”

— Jake the Dog, Adventure Time

Please believe in yourself… I believe in you!

I hold absolutely no judgment towards anyone who uses AI as an editor. You’re probably not letting AI write for you, you’re just looking for some feedback, and I totally get that. Again, I’d be a hypocrite if I pretended otherwise. If you read some of my older material, you smell some of Chat’s stink in those articles too.

Tools have their intended uses, but they also have inherent downsides. A hammer is a very specifically designed tool meant to insert and remove nails, but that hasn’t stopped it from being used repeatedly as a murder weapon.

All I ask is that you use AI with intent and trust your own voice. And, you know, pay attention to the way it talks. If you’re a writer, you can probably recognize phraseology and if not, this is a great opportunity to learn.

You are absolutely not too new or fresh or inexperienced or anything like that. You are an artist, regardless of your skill level. You don’t need to prove yourself to anyone, including me.

Trust yourself. Trust your voice. Trust your words.

And now, just to unnecessarily drive this point home, I’m gonna plug this article into ChatGPT with the prompt “go full ham using your voice to tighten the phrasing. Do your generic best,” so you have an example of what I’m talking about…

[note: I found the MIT article after I drafted this, hence it being missing from Chat’s version]

Thanks for reading, and stay balanced, my friends! ❤️ 🐻

PS - If I clock AI in your comments, I actually don’t mind because I mostly see it used to defend people and AI is really good at calling out incels and their ilk, so if that’s what you use it for, I’m actually all on board.

---

Your Voice Matters More Than AI’s Edits Ever Could

Before I dive into the nuances of AI’s pros and cons (that’s coming in the next piece), I need to get this off my chest.

If you’re a writer using AI as your line-checker or editor, please—please—don’t automatically accept every phrasing suggestion it offers.

Why?

Because those suggestions are the true giveaways. And what they give away is this: the homogenization of voice.

How I Know You Used ChatGPT

I started using ChatGPT earlier this year. A friend suggested it when I was struggling with tech issues. Within minutes, it helped me resurrect an external hard drive that had been dead for months. Years of photos—recovered in seconds. It was genuinely impressive.

Naturally, I started exploring its potential.

But I approached it the way I approach anything with stigma: cautiously, intentionally, and with an eye on ethics.

And as a professional editor who’s spent half a year talking with ChatGPT, I can tell instantly when someone’s using it to "clean up" their work. Not because of em-dashes—we’ve all caught on to that one—but because of phrasing.

There are patterns. Familiar cadences. Specific metaphors. Once you spot them, you can’t unsee them.

Even when the topic is meaningful. Even when the writer is someone I adore. That uniformity creeps in and smooths over what made their voice shine.

AI Is a Tool, Not a Voice

AI is useful. Genuinely. Especially if you’re writing without access to human editors. It can help you spot logical gaps. It can flag inconsistencies. It can help with structure.

But it also flattens.

AI, like a crutch, is valuable when you need support—but leaning on it too long weakens the muscle of your own voice. And that’s what I’m seeing more and more: writers who are afraid of being vulnerable, misunderstood, or seen as messy, turning to AI for reassurance.

The result? Work that sounds polished, yes—but also eerily similar to hundreds of other pieces.

That’s Not Weakness. It’s the Algorithm.

Here are some familiar turns of phrase you’ve probably read (or written):

• "That’s not [X], it’s [Y] dressed as [Z]."

• "It’s less about [X], more about [Y]."

• "[Adjective], but not the way you think."

• "Let’s unpack that."

• "Messy—in the right way."

They’re elegant. Rhythmic. Poised.

And they’re everywhere.

So are these overused emotional buzzwords: tender, sacred, unravel, holding space, cracked open, grief as teacher, permission to feel, radical softness.

Or overly tidy triads like:

"Not to be silenced, not to be shamed, not to be made palatable."

These phrases can be beautiful, but when I see them repeatedly across Substack, Medium, and newsletters, I no longer hear you. I hear Chat.

Learn From AI—Don’t Copy It

Here’s the thing. I use AI too. I run drafts through it to see what it catches. I ask it to point out weak spots in structure or logic. But when it tells me to change a sentence that already sounds like me?

I ignore it.

Because writing isn’t about perfection. It’s about connection. It’s about that little spark in the phrasing that no algorithm could ever recreate.

Your tangents? Your wordiness? Your sharp edges? They matter. They’re the fingerprint of your mind on the page.

Practice Is How You Get Better

I’ve been writing since 1999. Editing since 2006. Leading a media company since 2012. Self-publishing since 2023. I didn’t become confident overnight. It took years of missteps, messy drafts, awkward phrasing, and writing Final Fantasy fanfic that never saw the light of day.

There’s an old saying (that I’m probably misquoting):

Don’t expect Bruce Lee results if you’re not willing to put in Bruce Lee effort.

Writing is no different.

You don’t need to write like me. Or anyone else. You just need to keep writing—awkwardly, beautifully, honestly—until your voice sharpens into something unmistakable.

And if you need a crutch for now? Fine. Use it. But know when to put it down.

Because your voice? It’s enough.

I believe in you.

❤️🐻

---

Author’s Note: If you enjoyed this take on AI, consider checking out this sample of my novella series, The Vitmar Chronicles, where all of my balanced morals take the forefront in the worldbuilding! Volume II is set for release on August 22nd, 2025!