You just read three sentences that cost me $0.02 and made your brain quietly scream.

I've been noticing this pattern everywhere lately. Open LinkedIn. Scroll for thirty seconds. Count how many posts open with "X isn't Y. It's Z."

So I asked Gemini what to call it.

Turns out it's called negative-positive restatement. Rhetoricians call it "correctio." Copywriters call it "reframing."

I call it yet another AI tell.

The Robot Learned to Sound Important

Here's what's fascinating: this structure is technically effective. It creates a pattern interrupt. Validates the reader. Builds authority. Places emphasis in the stress position at the end of the sentence.

AI models trained on millions of LinkedIn posts learned that this is how experts sound when they want to be profound.

So now they use it everywhere.

Along with em dashes inserted into every paragraph—like this—to sound breathlessly insightful. And "not only... but also" constructions. And opening with "In a world where..." And closing with "Ultimately..." followed by a summary you already read.

The structure works. That's why it became a tell.

It's too perfect. Too polished. Too consistent.

Humans bury the lead. Use sentence fragments. Sometimes we just stop. At other times we just ramble with deep thoughts and profound insights that we think will stop the scroll of even the strongest focusers among us all while we completely forgot that the grammatical guide popularly referred to as punctuation even exists so readers can comprehend what we're even trying to say without having to carefully dissect a paragraph three and three-quarters times to either realize what the writer sought to convey or give up bewildered

Yikes! That was… well, let's just get back to the correctio.

AI Has Developed a Behavioral Trend

This is the part that actually interests me.

AI is sometimes talked about like it's this alien intelligence doing incomprehensible things. But here it is, exhibiting the same behavioral pattern humans do: converging on what works, repeating it, creating a recognizable dialect.

It learned from us. Now it's giving it back to us at scale.

Your teenage kid picks up phrases from TikTok. Your colleague starts every email with "I hope this finds you well." Regional accents develop because humans copy what sounds right in their environment. And don't even get me started on the shamefully overused "We apologize for any inconvenience." (Please unlearn that one!)

AI just does trends faster. And way more obviously.

What I'm Doing About It

I now have a "banned phrases" list for my AI draft writing team. No negative-positive restatements. No em dashes. No "in today's landscape." No "ultimately."

Not because these are bad—I like em dashes. Because everyone's AI is using them, and my content needs to sound like me, not like a statistical average of business thought leaders.

Hopefully you, the reader, will recognize that despite my use of tools to get things done, I'm still behind the post. I decided the message was necessary to share as I help others across industries in figuring out everything about AI other than how to spell it. I sought to address a pain point, a confusion, an opportunity that I don't want you to miss being aware of. I have perspective and experience that I know others find valuable. You do, too, and people need to see it, hear you, and experience the future potential you have to offer them… even if you work smarter not harder to do that.

An Idea That Never Sees the Light of Day Is a Wasted Gift

For many years, I painfully and slowly wrote to share value on my own. My content was verbose, not well-structured, or in most cases was never produced and shared at all because I couldn't carve out the time among many demands.

AI affords me the ability to help more people that I could solo. My hope is that the value being provided matters more than the nuances of the structure with which it is presented even if there is a tell that AI played a part.

(Also: no, I didn't write all of this myself. I dialogued with Gemini about it. Had Perplexity create a draft. Puked on it with feedback of 'not funny enough, do it again". Still changed it to my liking… maybe funnier if I'm lucky. Also fought with AI about three em dashes and two "it's not X, it's Y" structures. Old habits die hard for robots.)