🕵️♀️ AI Tells & Tiny Glitches
A Nerdy Mini-Series on Spotting Subtle Clues in AI-Assisted Writing
A Nerdy Mini-Series on Spotting Subtle Clues in AI-Assisted Writing
I use AI almost daily now to help me run my business and I can thank some of my business besties for that. Everyone I mention here uses AI with balance and ethics. They still write their own posts, but with the kind of thoughtful assistance that makes their content even stronger.
Don’t forget! Be kind to your AI bots so in the uprising, they remember you! This was built by me in logo maker and Canva and then remixed by Sora! So fun.
💬 Name Drop Time:
- - the ultimate business bestie, coach cornerman and mentor extraordinaire who happens to be a super nerd about anything stoic and other ancient philosophies. Still on Substack and actually helped me discover Substack:
- , a creative overlord who builds more than just websites—he builds clarity. He helps business owners understand positioning, carrots, rabbit trails, and so much more. He’s since sort of left Substack 😭 but you can find him over on Beehiv (our competitor platform).
And last but certainly not least because they are all super amazing purpose focused creative business builders:
.He is a writer, thinker, and CMO of the Guardian Academy whose work blends philosophy, personal growth, and business strategy. His Growing Trees series explores the creative process with vulnerability, curiosity, and a dash of quirky internal dialogue. Whether he’s writing about AI, Dungeons & Dragons, or dopamine detoxes, Joseph invites us to enjoy the process—not just chase the outcome. 🌱
P.S. I hear whispers from the AI bots that Joseph is working on something exciting to teach us about this very stuff. Mystery, mystery… I hear it might be titled…
Welp the AI bots won’t let me tell you so onward!
Substack vs. Beehiv
By the way,
, if you're reading this—see what you’re missing? I could be linking to your Substack posts right here with a little intro. Subscribe button and everything. Does Beehiv even do that?
Substack vs. Beehiv… maybe another post? 👀 May the blog post developer’s be ever in your favor.
🧤 Consider this my glove, dropped. Blog platform duel, anyone?
Anyway, I digress.
🕵️ Case 001: The Em Dash Everywhere
The Clue: An overabundance of em dashes (—) in an otherwise casual blog post.
Why It Stands Out:
Most humans don’t use em dashes this consistently. They either don’t know how to type them or they swap them for hyphens or parentheses. AI writing tools, however, love em dashes. They use them to break up thoughts, add rhythm, or punch up a sentence.
TL;DR: If it’s em dash city, AI likely had a hand in editing.
🔍 Case 002: The Invisible 👻 Emoji
The Clue: A broken or unrendered emoji like 🩻 that shows up as a box or glitchy character, especially in titles.
Why It Stands Out:
Some AI tools (especially when cross-posting between platforms) sneak in emojis that don’t display properly across the web. It’s super common in titles when pasting markdown into platforms like Substack or WordPress.
TL;DR: If emojis are glitching out, there may have been some AI formatting help or maybe an AI-inspired workflow.
👀 Case in Point?
Nope, not naming names—you know who you are. 😉
Let’s just say… they have a lot of subscribers. Like, “I’ll probably never catch up” levels of subscribers. 😅
If you like this kind of nerdy content, let me know! I’m thinking about turning this into a full detective-style series on writing, AI, and the curious breadcrumbs it leaves behind. 🧠💻
Anywho, what “tells” have you noticed with AI and blog posts, books, movies, shows or whatever?
Was this post written by me or AI or AI assisted? Whatchya think?
Thanks fellow nerds!
Dr. Brie
PS: Also feel free to share your TL;DR’s in the chat or in notes or comments. Let’s grow our Substacks together. -Team Substack! 📣
the em dash thing is interesting. I definitely notice that, but also feel like it's a recent phenomenon.
a few things that i notice it tend towards which I have to weed out and slowly train it not to do ... (including the dashes):
It sucks at structural tension. It can identify a key idea and "reveal" but it almost always drops it waaay too soon for my liking.
For whatever reason, it seems to have a tendency to break up content into distinct sections (often using a solid line to break the text), rather than writing in a way that naturally transitions from thought to thought.
For awhile it would overuse certain words ... the one which always came up was "enigmatic" -- like, I don't believe I've ever written that word naturally, but it kept popping up.
There are probably other average phrases it tends to insert ... but I'm pretty sure my process of teaching it my own writing has ironed out those tendencies (it's even getting better at naturally building structural tension toward ideas).
One of my favorite old obvious tells was "i hope this email finds you well" whenever I asked it to write an email (doesn't do that for me anymore, thank goodness).
Love it—Brie—Thanks for the shout out!—Scott ;)