
We all have “pet” words or phrases that slide into our prose like a favorite shoe that’s shaped to our foot and supports our gait—literary techniques that feel like our voice on the page.
Some writers come across more casual, using sentence fragments or words like “ain’t” and “c’mon!” Others turn to ellipses for hesitation and pop a statement (or question) in parentheses for humor.
Guess who’s been paying attention to how we express ourselves?
Yeah, LLMs slurped up our online writing and got a taste for the way we write. Their favorite words and phrases draw from our favorite words and phrases, and now they’re repeating them ad nauseam, shoving them into everyone’s drafts so people leaning on AI for their writing start to sound suspiciously similar.
What’s especially frustrating is that these are perfectly good writing techniques, and now that I see them spilling into my inbox, I hesitate using them.
In the hands of skilled writers, a light touch with these could produce a pleasant flow, but AI doesn’t yet seem to fully sense the need for restraint, balance, and pacing. It stuffs otherwise effective wording into their drafts like too many prizes in a box of Cracker Jack. They stop being special after the fifth appearance.
Once I point them out, you’ll start to spot them in blog posts, articles, newsletters, and social media. Here’s a small sampling:
1. It’s not X. It’s Y.
You’ll see variations of this contrast construction everywhere. Sometimes they’ll change up the punctuation, but it’s the same effect:
- You don’t need more ideas. You need movement.
- It’s not about having more time; it’s about using the time you have more efficiently.
- They didn’t abandon what already works. They built on it.
- You don’t need more examples. You need to watch for this technique and enjoy a chuckle. (That’s my addition; the others were from actual newsletters.)
2. Quietly (Quiet and quieter, too)
This word is such a favorite with AI, the minute I spot some variation of “quiet,” I question who authored the piece. Is everything unfolding as quietly as we’re told? I doubt it.
For evidence of its ubiquity, I did a search on the word “quietly” in my inbox and you wouldn’t believe how many newsletters turned up with some version of quiet (but mostly “quietly”). Here are some examples:
- [Learn] what’s quietly costing you conversions.
- [It’s] work that quietly saves companies a ton of time.
- I’ve discovered a quieter kind of ambition.
- I spent an entire year building this program. Quietly. Carefully. (Further down in the same newsletter…) I wanted you to hear it from me first before it quietly changes.
- [Those] projects quietly drain your team’s time, energy, and focus.
3. Set-up question (ex: “And honestly?” “Can I be honest?”)
You’ll see variations on this, too, but does this mean the author hasn’t been honest until now? Here are some examples from recent newsletters:
- And honestly? It should have been there from day one.
- Honestly? This is one of those episodes where you pause, rewind, and take notes.
- It’s a weekend… and honestly? It’s one of my favorite times to work.
- Want to know the truth?
- (Though not in question format, this is similar): Because here’s the truth.
4. Rule of 3
This is a solid writing technique I’ve taught in The Art & Craft of Writing (ACW) that AI shoves into nearly every paragraph, sometimes back to back. Because it can produce such pleasant pacing in the right hands, I hate to bring it up. In fact, you’ll find it in my own work—maybe in this very blog post.
Nevertheless, here are examples pulled from my inbox:
- A book that becomes a workshop. A series of posts that becomes a live gathering. An idea that becomes a space people can step into.
- And you haven’t yet had that breakthrough moment, even though you’ve done the work, even though you’ve got the knowledge, even though you’ve got happy clients and testimonials.
- (From an earlier example): This is one of those episodes where you pause, rewind, and take notes.
Again, as with all of these, the rule of three can form the basis of quality writing, but AI has so enthusiastically included them in their drafts, the result is often clunky, obvious, and annoying.
The last example is kind of painful for me.
5. My beloved em dash
I use the em dash often, as it’s the most versatile punctuation at our disposal.
It connects ideas visually better than a semicolon or comma. It adds energy, urgency, or hesitation, depending on context. And it breaks up long sentences with a natural pause—no need to add a period and divide a long sentence into two choppy independent clauses when you can insert an em dash!
The problem is that AI writing programs use so many, you’d think they were crushing on Emily Dickinson. At a glance, AI drafts appear sewn together by a beginning seamstress, sentences stitched and held by multiple em dashes throughout every paragraph.
But I use them abundantly, too.
I’m not the only writer clutching her beloved em dash, refusing to give it up. Ann Handley has been rooting for the em dash on behalf of writers everywhere, as well. She argues:
Large language models are trained on millions of human-written sentences. (Mostly stolen, it must be said.)
And we modern online writers use em dashes a lot. They mimic our conversational, emotional, sometimes delightfully digressive voice—the kind that wanders, like it’s doing right here—before it then plops into a seat and parks itself.
So when AI learned to “sound human,” it picked up our favorite tic: the em dash.1
Choose and Use What Feels Like You
I unite with Ann Handley and say go ahead and use the em dash if you want.
In fact, as you start to notice how much AI is copying us (but can’t achieve human finesse…yet), you may feel confident reclaiming contrast construction, set-up questions, and the rule of three, as well.
Let’s incorporate them into our work when appropriate—without worrying we’ll be accused of using AI in our creative process.
Because soon enough, the robot writers will shift to a new set of pet words and this list will be irrelevant.
I don’t want to sound like an AI writing app, but if I choose effective techniques and tropes to add the flow I want for my writing and they happen to be popular with AI…so be it.
Ross O’Lochlainn ends his emails with what he calls a “100% Human Guarantee.” He adds the image of a thumbprint alongside this:
This email was lovingly crafted by a human for a human. While I may use AI during my creative process (brainstorming, images etc), I will never use an AI or robot to represent me. It will always be me you see and read.2
Do we have to actually say this? Are we finally at the point where we must pledge that our projects are entirely human-produced?
I used to think a statement like that would be overkill, but honestly? Now I kind of like it.
So if you’re wondering, this blog post was written entirely by me, for you.
And I didn’t use AI during the creative process, either, so all of my AI mimicry is to make you laugh.
What AI words and phrases have you been noticing?
Resources:
- Whose Voice Is on Your Pages? Human vs. AI Writing
- Write to Discover Your Voice
- Metaphor Magic: Wave Your Pen Like a Wand
- To Be More Creative, Write a Letter to Your Reader
You want to write your truth (without AI)
Find your spark and hone your voice with guidance from a human writing coach

You’re ready to tell your story and write with confidence (without sounding like you’ve copied output from an AI chatbot)? Work with Ann Kroeker as your writing coach in her signature group program, Book-Ready Writers. You’ll sharpen your writing skills, gain confidence, and get the human input you’re looking for. Your spark is real. Let’s fan it into finished projects ready to publish.
Footnotes:
- Ann Handley. “The Em Dash Is NOT an AI Tell: Justice for the Em Dash! – Ann Handley.” Ann Handley, 28 Nov. 2025, annhandley.com/em-dash.
- Ross O’Lochlainn. Newsletter with subheading “The fallacy of ‘Selling too hard'” emailed to Ann Kroeker. 25 Feb. 2026.
