You can’t make this up–but AI can!
AI is everywhere now—from writing advice to government reports—and like most tools, it’s only as sharp (or as silly) as the person wielding it. I’ve had a few moments lately that made me laugh out loud, shake my head, or just marvel at how weird things can get.
Here are three quick snapshots from the intersection of authorship, AI, and absurdity.
1. When AI Called Me Out for Being Ahead of My Time
I pulled a 2017 manuscript, *The Empath Gene*, out of “storage.” I wrote it in about 2 months after I got home from Ireland, but I never had all the chapter files in one place until now because of moving and not being able to find 2 critical scenes.
In a final round of polishing before sending it to my editor, I ran it through an LLM with a list of “common AI-isms” to flag for me so the folks with pitchforks don’t spot one and harass me. Needless to say, I was widening eyes and quickening pulses in 2017—and a lot earlier.
And no, I’m not giving up my em dashes, ellipses, or Oxford commas or whatever a non-English major doesn’t think is normal for an author to use. That said, using AI to help me find AI-isms was a hoot.
2. When AI Wanted a Blizzard in Florida
In the 2025 manuscript I just finished, *The Book of Heroes*, I’m prepping it to go to my editor. I dropped it into an LLM to tell me if I had any plot holes or inconsistencies. I’ve done this before and usually nothing it tells me is truly an issue, but AI tools freak out over time travel and parallel universes, which are by their nature inconsistent. I’ll usually get about 10 flags and 9 to 10 are nothing.
But this time, I got a flag that made me laugh: the book takes place in February and my characters aren’t struggling with heavy coats.
Well, duh. It’s set in central Florida—we don’t OWN heavy coats.
3. When AI Cited the Twilight Zone
Recently, I’ve seen several Government reports that cite studies and other reports that don’t exist. Hard for me to trust any Government office that puts out an official response that was made up.
And I know exactly what they did: they prompted a particular tool (I suspect I know which one because people who do this aren’t even aware of deep research tools) to create a report, with citations, that backs up their point of view. Generative AI, being generative, generated a report. Aka, made up stuff.
No one could be bothered to *check the citations* to make sure they were real. Rookie mistake.
Sometimes AI makes stuff up. Sometimes the government forgets to check. Either way, it’s still my job to ask: Says who?
Three AI fails. One amused author. Zero regrets.
A Southern witch returns home. Secrets won’t stay buried. A chance to confront and heal—or face the consequences.
Visit the Book Page →