Skeptic or Geese?

Mark Honeychurch - 26 May 2025

In my efforts to wrestle with the push to include AI content in our newsletter, I’ve been chatting recently with our newest regular contributor, Patrick Medlicott, about why I prefer content that is written by humans. One new reason that popped into my head is that I know, from editing people’s stories every fortnight, that each of our writers leaves their own distinct fingerprints on each and every article they write. Bronwyn, Katrina and Patrick each have their own tell-tale signs that they have written an article - certain repeated typos, punctuation choices, use (or lack of use) of commas, sentence length, repeatedly used words and phrases, etc. Each of these things has become very familiar to me, and through my editing I try to clean up each article without losing too much of these fingerprints.

As I said to Patrick the other day when I’d received a piece from him that had originally been written by AI, but that I asked him to resubmit after writing it in his own words:

“It’s quite funny that as an editor I would want to see typos, but in this day and age they’re a telltale sign that something probably hasn’t been written by AI or copy/pasted from another source. I’d much rather tidy up your writing and have something we can publish that’s in your voice than publish content from AI, via you, that lacks any identifying style and feel. I think each of us regular contributors to the journal has a distinct writing style, and I assume that most of our readers will have become accustomed to our idiosyncrasies and turn of phrase.”

I know that, like many other reasons to not use AI for writing articles, it’s likely to become redundant over time as people learn to write prompts that coax an LLM into writing in a distinct style, or even mimics their own style - but, at least for now, we’re in the golden age where I think I’m able to spot AI-written content pretty easily. Of course, it might be that I’m not detecting it all like I think I am, and that there’s well-written AI content that’s flying under my radar. But, if that’s the case, then well done for fooling me, and not so well done for ignoring our rules on submitting articles.

In this week’s newsletter, following the AI theme, I grapple with Google’s AI and end up with some pretty disappointing results when prompting it for answers to some skeptical questions (as well as some funny attempts where the AI tries to explain the unexplainable. But, before that, Bronwyn has written an update on the Two by Twos, looking at what’s happened to them since she last wrote about them two years ago. Katrina’s written about mansplaining, and tries to see what the scientific literature has to say about it (oh, god, I hope I’m not mansplaining her article right now!). And finally I’ve watched, and re-watched, the Under His Command documentary series about Destiny church, and try to review each episode - something I’m sure we’ll have more to say about on our podcast, as both Katrina and Bronwyn have also watched it.