5 Ways I Spot AI Use In My Own Writing 

The writing features em dashes

 

When evaluating my own writing, em dashes are always a dead giveaway that I’ve used some form of AI. Prior to these tools flooding the market, I didn’t know what an em dash—or even a regular das—was. Punctuation usage in general is an obvious indicator that I’ve used AI to craft copy. Whenever I used to write back in the dark ages, mostly my fanfiction, I did so in long rambling blocks of texts—just one breathless idea with a slew of missing apostrophes, commas, and periods.


The sentences are all the same length

 

As previously mentioned, if my writing has clearly defined sentences with periods, that’s a red flag that I’ve used AI. With that said, there used to be times where I would be in an absolute flowstate, jamming my fingers into the keyboard with such reckless abandon that I would accidentally punctuate sentences. Typically, this would occur during the more passionate moments of my fanfiction, of which there were many.

 

These days, if I’m reading my work, it contains sentences, and those sentences are all uniform in length then I can confidently assume that I’ve used AI. That’s a level of control in my “Boston Legal” fanfiction that I was never able to previously achieve.


The text is needlessly complicated

 

If my writing feels unnecessary wordy, using 10 words to say something that could be said in a few, I know I’ve used AI because I only know 20 words total.

 

In truth, some of those words are actually character names from “Boston Legal” (Tara Wilson, Denny Crane, Alan Shore), so the actual number might be closer to ten—which is why my work always revolved around the same plotline (William Shatner’s Denny Crane forgot to wear pants in the courtroom). If my writing is overly verbose and uses words like ‘verbose,’ which I don’t know, then it was probably written by AI.


The work lacks real emotion

 

If my writing lacks real, authentic emotion, then I know it must have been written by AI because when I used to type up my “Boston Legal” stories, they were incredibly steamy. While I only knew a handful of words, most of them were provocative in nature and meant to elicit distinct feelings of passion. We’re talking unpunctuated, uncomplicated courtroom hijinks that dispensed of the details in favor of smut. My AI output, while much more coherent, feels perfunctory and soulless by comparison.


The writing uses data that is factually incorrect

 

If my erotic “Boston Legal” fanfiction uses figures and data that are clearly incorrect, you can assume it was written by AI. Between the sensual moments, I always aim for verisimilitude by citing actual cases. Something like, “Alan looked sternly at the judge and said, ‘According to 1962’s Miranda vs. Arizona, you had better give me a smooch, baby,’” which is obviously very articulate and sexy, but provides the wrong year. That landmark case occurred in 1966.

 

With AI, facts and figures are often pulled from unreliable sources or made up out of thin air. A piece of well-known data, like William Shatner’s bicep measuring at 7.9 inches can somehow be mistaken for being 11 inches, which is—of course—ridiculous. When major gaffes like that occur in my work, I can be fairly certain that it was generated by artificial intelligence.

 

When it comes to AI, this is uncharted territory for all of us. As a people when presented with the unknown, fear is often the first reaction. And yet, if these tools can be used to haphazardly draft seedy fanfiction then imagine what other possibilities they might hold.

 

There’s one thing I know for sure: As generative AI becomes more and more sophisticated, it’s only going to become easier and easier for me to spot AI in my own writing.