Is AI your sparring partner or your ghostwriter?
Ghostwriter caricature created in Midjourney
This singular difference determines the value of your voice and the usefulness of your message. Today I want to talk about losing your voice to AI, which seems to be a trend I've been noticing. I love using AI to assist me as you know, but I truly and wholeheartedly believe we need to learn how to use AI effectively to stay authentic while doing so. Bear with me, this article is for your personal learning growth to help you sound uniquely you next time you guide the prompt conversation, especially if you have fallen victim to AI ghostwriting for presentations, for social media content, your emails, and maybe even the anniversary card to your partner.
What I love about being human is: we have senses and a radar to pick up on artificial writing. No matter how good the generated content seems to be, we remain savvy and instinctive at spotting the intruder.
This works at age 17, age 28, age 45, age 59, and over 78. Here's an infographic example shared by NP Digital on how good we are at spotting AI content: that's 82-88% depending on the generation! That's a high percentage of human instinct kicking in to spot the AI ghostwriter. In my opinion, far too high, to want to risk losing your voice and those who enjoy reading your words.
Source: Search Engine Journal, NP digital
The ugly side-effect of lazy usage of AI is, that readers will stop reading your content as soon as they realize it was the artificial hand of your AI ghostwriter and not your own voice. People want to hear what YOU have to say, not what your preferred LLM has to say. Take this with a pinch of salt and next time you copy-paste your output directly, think of these few tricks before you post generic content:
Ask yourself:
Have you provided original thought to your writing input?
Have you supplied multiple original samples of your writing style?
Have you tweaked and reported back (to the LLM) what doesn't sit well with the writing tone of voice that it supplied you with?
Have you instructed a set of output requirements that side with your writing intention and tone of voice?
Have you added your style of punctuation and sentence rhythm or did you let ChatGPT or Gemini override your writing input, or worse, completely write it for you?
Did you give attention and detail to how you briefed your content to sound uniquely you?
Have you been reporting back with writing sample corrections? Yes, the good old teaching method of correcting a text is necessary.
And who did the final editing touches? Was it your critical eye or just output by an LLM?
I teach AI principles and thought processes, which also means I read AI content all day long. I can assure you, I will know if something was written by or corrected by an LLM. It doesn't matter which LLM it was, they all have tell-tales. But what matters is that the writing you let it assist you with still reflects your voice. So in a world where AI is seen as the solution to time saving and the tools are more accessible and usable to all generations, we are seeing a lot more content being written and directly posted or shared without final editing and alignment of your voice.
It almost feels as though we are losing the originality of our voice, which remember: is only uniquely yours. But if we are a nation of LLM users, who have all started artificially writing, the result is that we all share the same voice. And correct me if I'm wrong, no one likes to sound like 3 million other people. Right?
Naturally humans will not want to continue reading, once they realize, it’s the same voice again and again and again.
What’s missing is the nuance, the humanity, the raw, unfiltered mark of someone’s mind at work. We need to stay true to our voice, and we do need to put in effort to assess how it sounds. So here is my tip for your next piece of writing, when you prompt an LLM. Whether for research or for proof-reading your wording and writing, make sure it still sounds uniquely yours. And yes, that requires more effort. But you will sound authentically you, bringing more value to your community and readers.
When working in tandem with ChatGPT, Gemini, Claude, or other LLMs:
Give it a sample of your writing style.
Instruct it how you like to structure and format your writing.
Report back with critical feedback, if it doesn't feel like your writing, then explain what would feel true to your voice. You need to teach it, like you would teach your real ghostwriter or your real personal assistant.
If you are not infusing your own voice and tone, you are letting the LLM water down your content. And humans will know. We sense it.
Telltales to avoid, in order not to sound generically artificial:
Avoid having emojis all over your writing. No one does that naturally. Especially not with rockets, arrows, sparkles and pink flamingos in a professional sounding context. Before ChatGPT no one in their right mind would have added these emojis into a professional presentation for a client.
Avoid long dashes all together. What happened to good old punctuation? Commas and full stops and semicolons are as professional as you need it to be. That's what worked in school, why wouldn't that be academic enough for articles and posts in 2025? Simple sentences always work best.
Avoid comparison sentences when they are absolutely not necessary. If you are comparing facts, that's an effective way to make a stance. Keep it. But, if you have a sentence provided by an LLM that reads like this: "The problem isn’t the [noun]. It’s the way people [verb] it."; then please rewrite it. The thing is, if you read a sentence like this once, it sounds great. But if you are reading 100 content samples all written by the same hand aka ChatGPT, Gemini or any other LLM, it sounds artificial to read these types of sentences repeatedly. Because that is part of the scripted voice of LLM outputs. Great once, but terribly repetitive.
PS: Could you tell that I wrote this article without AI?
I'm assuming you did. Sorry about the lengthiness, I love writing, which is why [Warning!] I am very picky when I spot lazy writers, who only copy and paste LLM text output.
I hope this excites you to try and personalize your next set of content! Stay tuned for more prompt output analysis. I feel like this is a very necessary subject to learn about, now that more and more people are approaching it as a very resourceful time-saving tool, which it absolutely is! :)
Also, side-note to all the authors out there: Your unique style and writing voice will never seize importance. If anything, I hope that's what you gain from reading this. Being human and unique matters even more now in the age of AI, whether you write with assistance or without assistance.
👉 Interested in learning how to control your creative process with AI? Join the course waitlist here.
