How to Avoid Looking Like You (Need To) Use AI

I'm confused. Aren't we all supposed to be using AI?

That's what I thought. But if that's true, then why are people so desperate to hide that they use it? And why do those same people rejoice whenever someone else is caught letting an em dash slide into their prose?

I've been giving this some thought in between daydreams where I delete my LinkedIn profile (bliss!) and I've come up with what I think are the rules everyone plays by but nobody explains.

Let's see if I get this right:

First, you should DEFINITELY know about AI. You should have strong opinions about all the different services. ALL OF THEM! Which one is best for helping you with your molecular cell research or whatever? (Perplexity.) Which one is best for inserting yourself into Three's Company episodes? (Runway.) Which one is most likely to use all your data to try and sell you Ozempic in six months? (All of them.) Casually mention you used Gemini to perform some task that EVERYONE KNOWS Claude would've done better and you will reveal yourself as a putz with a landline.

OK, so know all about AI. Duh. BUT — and this is important — NEVER admit to using it. You may think it's OK to tell your coworkers, because, hey, everybody does it right? Well, yes, everybody uses it, but admitting to it is another story. For writers, this is especially true. Putting something out there that has the whiff of AI is the last thing you wanna do.

But don't go canceling your ChatGPT subscription yet because there's a hierarchy. Using AI for research is fine. Using it for editing? Eh, a little dicey. But using it to actually write something from scratch is to court mockery. And God help you if you use AI for a cover letter. I'm told the bots that read and analyze cover letters disapprove of AI writing. It seems AI is learning the very human act of self-loathing. Oh, AI — we'll make a writer out of you yet!

There's also the matter of proof. Since everyone's apparently detecting AI by vibes alone, a new tell has emerged: the deliberate imperfection. A run-on sentence, a stray "lol," an em dash swapped for a comma splice — you name it. Nothing screams "certified human" like willfully leaving in a typo. Images flip that logic, though. To err is human. To add a third arm is machine.

So, are you supposed to be using AI? Sure, as long as nobody can prove it. You always want to maintain plausible deniesability.