I saw something sad and disturbing on Threads this morning. A writer shared a 2-star review they received via a screenshot of it, and in the review, the reader accused the author of using AI because instead of referring to the characters as “mother” and “father” or “king” and “queen” in certain instances, the author referred to the characters by their names.

That’s it. That was the complaint.

Umm…WTF?

I wrote in response to the author, “I’m sorry this happened to you. I wish there was a solution to this. Accusing authors of using AI just because they disagree with how the book was written is low-hanging fruit.”

I’ve seen a lot of authors on social media relate stories of being accused of using AI, when they themselves hate generative AI and would never use it. It’s becoming a witch hunt at this point.

The thing is, as I wrote in response to the author, I don’t have a clear solution to this problem. If someone accused me of using AI, I could only prove I didn’t use it with The Name and the Key, because I have several different files that were created over the course of writing it that show it dated from 2024 all the way to 2026. NOt to mention emails of my thesis to my mentors and crit partners in 2010-2013.

I would’ve been able to prove originality with Son of the Siren, except once it got published I deleted the file drafts (except for what the publisher looked at) and I threw away my plotting notebook.

I wouldn’t be able to prove it with The Step and the Walk because I didn’t save multiple files while I wrote the novel. I just used one file. If there is a way to check file history and save history through Word, though, I’d be able to prove it (I just don’t know how to do that).

Authors shouldn’t have to show their method or progress or notes or notebooks to prove that they have been working on a project on their own for months or years without the aid or use of generative AI. But will it come to that?

We have to remember that generative AI was trained on thousands of stolen works, where a multitude of books, fanfics, short stories, blogs, and articles were uploaded to AI to train it how to “write.”

Authors don’t write like AI. AI is trying to write like authors.

While I’ve talked about some “tells” in AI–mostly when I’ve analyzed the emails I get soliciting book-related services–there really isn’t a clear cut way to prove AI was used unless the author stupidly leaves their prompts in the book (which has happened more than once) or just comes out and admits it.

You can’t rely on AI detectors (which are also often AI themselves) to determine whether or not a work was composed with AI because they are unreliable. Famously an excerpt of Mary Shelley’s Frankenstein was uploaded to an AI detector and was determined to 100% be AI-written. Oops.

Also, AI is sycophantic–it tends to support or affirm what the user believes or wants. It wouldn’t surprise me if someone fed a manuscript to ChatGPT and told it “I think this was written with AI, what do you think?” for the LLM to turn around and say “I agree,” even if it wasn’t.

While readers can say that they have credible reasons to claim a book was written by AI–and this has happened with authors like Mia Ballard and Jamir Nazir–should the author then come back with all of their prewriting, notes, drafts, and save histories to prove otherwise? (As a sidenote, I’ve watched YouTube deep dives into both authors’ works and while I can’t confidently conclude that they were AI-written, I will say the quality of the writing was poor based on the excerpts that were shared. Sometimes the writing just made no sense).

How does an author prove they never used AI? Should they have to?

And I think the bigger danger here is, like with the story above, readers accusing authors of using AI because they didn’t like the book or its writing. I think that would be every writer’s nightmare, to be honest. No one likes false accusations, or accusations made lightly that can potentially ruin careers.

As an adjunct professor I sometimes get vibes when I think a student uses AI, but because I’m not sure myself, I hold off on accusing them, because I’m afraid I’ll get it wrong unless they do something really obvious, like, for example, leave the prompts or instruction in.

All I can do is repeat myself: I don’t have answers to this. I don’t know how to stop readers falsely accusing authors of using AI, and I don’t know if the burden of proof should be on the accuser or the accused.

I just know that what I saw on Threads this morning really disgusted me and made me feel sad for the person who was accused of using AI, for an exceptionally stupid reason. I hope it never happens to me, and I hope it never happens to anyone else.