So many people these days, people I otherwise respect, use AI to write these days. Anyone who has read even a bit of AI writing can always tell when an article is machine-generated. Better to be a poor writer with a soul than a great one who is just a product of the machine (not that AI is anything close to great writing).1
— Robert Spencer
Parts of my writing process involve AI tools. Everything here is planned, edited, and signed off by me before it goes out. — EKO2
Okay, well then! Thank you, EKO. It’s comforting to know you’re looking at the AI-generated content and “signing off” on it before it invades our computers and smartphones. We can all go back to staring at our screens during breakfast, lunch and dinner—anytime really—knowing you’ve put the reins on Lord AI.
I’ve been reading EKO every now and then (mostly then) for quite a while and not once, that I can remember, has he ever disclosed his use of AI until Substack made AI detection available and that little message ONLY shows up if you do an AI check on his content. The post where this message appears flags it is 100% AI (see below). The detection software, Pangram, could be wrong, of course, but more and more I am not doubting its conclusions.
No AI detector is flawless, but Pangram is pretty much the gold standard. Their algorithms detect AI-generated text more than 95% of the time, with a false positive rate approaching zero. I was very happy to read about this feature, and I think more social media platforms should do something like this before their content descends into a morass of platitudinous computer-generated nonsense.3
I ran my writing through Pangram and it considered me—100% human. Good to know, I was afraid I might be a self-conscious AI in a malfunctioning Matrix. I ran a few other writers through who got the same 100% human score.
At least EKO admitted it and the tool confirms it. I have a couple of EKO’s digital books. I’ve noticed he’s been putting out a lot of books….almost too many. Probably too many. I’ve written some books, it’s not easy, nor is it a quick process unless one is publishing public domain content, something already written or using AI. No way could I put out the volume he does in the time he does. But I don’t see any mention of AI being used in these books. If he’s using it on his Substack, I suspect he’s doing the same in his books. Why not say so? The section I tested in his book Kennedy: Unsealed Archives V, was flagged as 100% AI.
I assume writers wouldn’t want to admit they’re using AI unless they had to. I have yet to see a writer admit it up front. Before the tool came out I accused a writer of using AI and if any of you are wondering—it did not go well.
I think EKO is trying to get out ahead of it…at least for anyone checking. I’ll be curious to see if some writers start writing defenses of their use of AI. It may be difficult to write something from scratch. After a few frustrating starts and several beers it’ll be a lot easier to simply write an AI prompt:
Defend my use of AI to write essays on my Substack
I suspect many people will not care if they’re reading AI. Eventually it will be so common to write with AI those not writing with it will be considered misinformation spreaders. Maybe that’s the eventual point. Maybe that’s the endgame: AI becomes the standard, human writing sinks to the bottom of a contaminated lake of AI slop.
I’ve already seen spam emails explaining to people how to publish full books on Amazon. “No writing or research required!” Abracadabra! “Authors” will learn to write great AI prompts for novels, poetry, screenplays and emails to Grandma. Then they’ll get the AI to write the prompts (Why fuss with that?), oh hell, finally they’ll let the AI do everything while sipping margaritas by the pool with bikini clad AI robots capable of riveting conversations and lots of other things.
I started going through people on Substack to run Pangram on them. Here’s some of what I found:
The BIG WILD MIND MACHINE responded to a comment of mine about his 100% AI score from Pangram. First he or she or it wrote this:
There’s something wonderfully recursive about responding to an essay on mediated reality by asking another algorithm to tell you what’s real. Whether the detector is accurate or not almost becomes secondary. The essay asks who gets to authenticate reality. The reply outsourced that question to a machine.
Then the entity responds again before I responded to the above:
One more thought. The detector has no access to the creative process. It only sees the finished artifact. It doesn’t know who conceived the ideas, who asked the questions, who rejected twenty drafts, who kept pushing back, or who decided when the piece finally said what it needed to say. In other words, it mistakes language generation for authorship. That’s an interesting assumption. The language model has no independent desire to write about the Fourth Estate. It doesn’t wake up wondering about propaganda, mediated reality, or the confidence game. Those questions came from somewhere else. So what exactly is “100%” measuring? It seems to assume the tool has agency while the human disappears. Ironically, that’s almost a perfect illustration of the essay itself: another algorithm being asked to authenticate reality without any access to how that reality was actually constructed.
All this really says is: I use AI to write, but it’s much harder than you think! And, of course, Pangram flagged that as 100% AI. Good grief! Dystopia arrived and I didn’t even notice. And if this big wild guy is rejecting 20 drafts, why not just write his own draft and spend that time editing it? Maybe the Big Wild Mind isn’t so big after all?
The day after this BIG WILD MIND MACHINE published this essay flagged as 100% AI: “AI POWERED BOT ARMIES ARE INVADING YOUR MIND.” Okay! Intentional or unintentional irony? Inquiring big wild human minds must know!
And later that day, it posted this: “MY ROBOT: INTELLIGENCE FOR NORMIES.” I thought it might be directed at some of my comments, so I read it. I liked it. I’m reluctant to admit that. Yup, flagged as 100% AI.
There are those, Big Wild Dude included, who believe that working with AI to create something is a new kind of writing, a new art form. An essay on his site, The New Artists Who Collaborate With Machines, elaborates on this idea and of course the first four paragraphs are flagged as 100% AI. I didn’t test the rest.
I hadn’t thought much about AI writing content until I realized an author I liked was using it to write. I am resistant to an AI writing content with an author. Probably, resistant is too mild. I am opposed to it. People might say, yeah, well people feared the influence of the car and telephone too.
But AI is like Frankenstein’s creature. It’s something brand new. With the phone, you’re still talking, with the car, you’re still traveling. With the AI—what are these writers doing? How much does the AI write? Whether you’re using AI to become famous, save mankind or write a Substack post, it should be approached with caution. We’re not even sure how it works.
As with anything this new and this powerful there will always be unintended consequences. I wonder what those might be for writers using it to write? Will it stunt their thinking? Will it subtly change their views? Will it literally rewire their brains? Will their ability to write and think be crushed by AI? Who knows? No one. A little humility and trepidation might be a good place to start.
Outsourcing your thinking to LLMs [Large Language Models] risks degrading one’s skills and making your work more impersonal. — Lincoln Michel4
My son thinks these writers are lazy and talentless, like artists who produce “modern art.” eugyppius, a Substack writer I follow, thinks the same. I wonder too, if many of these writers have little confidence in their own writing. After being initially impressed with first attempts at prose, new writers realize, hopefully, two things: they’re not good at it and that writing is hard.
AI makes it easy.
Why aren’t these authors acknowledging, up front, that they’re using AI to write content? I want to know. Just own it. If you think it’s okay, say so. Why the reluctance? Why so defensive?
All this is making me rethink my use of AI to generate images I want for memes. One of my subscribers (thanks, Dave) suggested I create them myself and I am considering that. Watch out, Inmates—it won’t be pretty—I am NOT an artist! But maybe I’ll try one (Okay, I tried two—as you can see. Maybe next time I’ll try hand-drawing them.).
In addition to EKO and BIG WILD MIND MACHINE there were some other writers I’ve suspected were using AI that got flagged using the AI detection tool. The most disheartening one for me was A.C. Rosenthal (I wrote a longer piece about that while also reviewing one of his books).
Rosenthal is a writer I have been highly recommending to readers of The Asylum. His writing is flagged as mostly AI with some human input. I’ve run two sections from his two books (The Two Muhammads and The God They Rejected Isn’t Real) through the online Pangram AI detection service. Both sections came up as 100% AI.
I reviewed The God They Rejected Isn’t Real. I liked the book. I’m a little embarrassed, now, that I did. Something about it feels wrong. This is a book about God and if partly or mostly written by an AI with no soul, no intent, no faith, no skin in the game—what have I just read? I’ve since had some exchanges with Rosenthal. Here’s what he said about his process:
I use AI the way I use spell-check. It’s a tool, not an author. It might hand me a skeleton from a pile of bones, or help me articulate a thought I’m circling. But the pile of bones is mine and I go through every word of it. Every paragraph, every sentence. I add, I cut, I rewrite, and I rewrite again, until the thing on the page is me. If a line doesn’t sound like me, I work it until it does or it goes in the bin. The tool never has the last word. I do. Always.5
Is that enough? And what’s a “skeleton” or a “pile of bones”? What does that mean? A whole AI written essay? A paragraph? A few phrases? What? What might A.C. have internalized without even knowing it? What came through the book that would not have if he’d not used AI? What would his writing be like without any AI influence, an influence no one fully understands this early on?
How do I even know that A.C. only writes his Rosenthal Substack? He has written that it’s not his real name. Maybe he’s some dude sitting in a bedroom at his parents house with six computer screens in front of him running 25 Substacks all written by AIs: Barbecuing for Dummies, Covid Sucked, Trump is an Asshole, Trump is God, Why you Should be a Christian, Why you should be an atheist, Why MAGA will Win, Why MAGA will Lose….I have not doubt I could do that if I wanted.
In my response to A.C. I wrote that I thought his writing would be better without this process. If he does that much editing, why would he even need an AI? “Capable authors,” eugyppius wrote, “will find it much easier simply to write their own goddamn blogposts.”6 Rosenthal’s explanatory paragraph above? It gets flagged as 100% AI. Long sigh—of course it does.
Will the human A.C. please stand up!
That’s what depressed me. Two of his long comments to my Substack pieces were flagged as AI. One 51%, the other 100%. It made me exhale slowly and shake my head—not in anger, but disappointment. He, it, she can’t even write a comment without AI? Who or what am I conversing with? Who or what have I been recommending to readers of The Asylum?
Is this transhumanism in its infancy? Instead of attempting to fend off the Borg, will humans voluntarily off load their intellects and talent to an AI? Probably. Likely. Yeah, they’ll do that.
A.C.’s comments on my essay7 and book review8 were flattering. Maybe a better word is magnanimous. Too much so. I guess A.C. (or the AI) is a better Christian than I am, but had he written the things about me that I did about him—I’d have shot back. Probably been defensive. Maybe reacted too quickly. Maybe regretted it. Who knows? But we humans are messy. We don’t always get it right.
Humans have opinions and emotions. AIs have algorithms.
Here’s the problem: once you suspect someone is using AI extensively to write, you start looking at anything they write as suspicious.
It looks like A.C. turned off AI detection on his latest posts (you can do this on your own posts in Substack if you want). But if writing being flagged as AI shows your competence as a writer, something A.C., or more likely the AI, told me (I’m not kidding), why turn it off? It should be honorable, not shameful.
As Scotty says in J.J. Abrams first Star Trek movie: “I like this ship. You know, it’s exciting.”
A few weeks ago I was talking to my son and wife about what I’d been reading on Substack. There are some phrases, paragraphs and whole passages out there that have a quality to them that sound a little odd, a little off. It’s a way of writing that feels not quite right. I first noticed it with Rosenthal because I was reading him almost every day and then began to notice it elsewhere. I can’t define it exactly, except to say these are phrased in ways that sound strange to me. Here are some examples:
You may find yourself agreeing with parts of this book.
You may find yourself offended by others.
Good.
The purpose of a mirror is not to flatter the viewer.
The purpose of a mirror is to reflect.
And this book contains a mirror large enough for an empire. — BIG WILD MIND MACHINE9
Here is something that should reset the playing field, because the playing field is tilted in your imagination in a way it is not tilted in reality. — A.C. Rosenthal10
I want to start with something most people never consider.
The Gospel of John was written last.
Not because John was an afterthought. Because John was a culmination. — Nabu11
If someone had just read that to me, I’d have guessed it was Rosenthal. Hands down. Hell, maybe it is.
I want to start with something small.
Not a big moment.
Not a dramatic turning point.
Just something quiet. — Deepshikha Chaudhari12
I want to start by agreeing with him. Not strategically. Actually. A belief you hold in a way that no possible evidence could ever dislodge is not a belief you arrived at by looking. — A.C. Rosenthal13
This way of writing has been irritating me. It’s more than the common phrases such as “I want to start with,” it is also what comes after it—phrases that sound…not quite right, unfamiliar, not human. The fourth example above doesn’t come from Substack. All these examples are from passages that were flagged as at least 90% AI written.
Or another sentence construction is something like this, “I need to begin here, because without this beginning, it will be impossible to end.” That’s the feeling of it, not the exact phrasing. This feeling, this thing I can’t quite describe is what started me down this overgrown, winding, muddy, slippery path. It’s like the scene from The Matrix where Morpheus says to Neo:
What you know you can’t explain, but you feel it. You’ve felt it your entire life, that there’s something wrong with the world. You don’t know what it is, but it’s there, like a splinter in your mind, driving you mad. — Morpheus, The Matrix
Yes, a splinter in my mind. It’s been bugging me for several months now and the new AI detection tool has taken me to and down the rabbit hole. It’s a deep one.
So I’m not going to read AI writers any longer. Maybe I’ll come back to them at some point, I don’t know, but I’m done for now with a problem I never had to consider for most of my life.
I’ll keep writing 100% human prose—for any Inmates who still care.
Time to start reading the classics again.
Download Meme
You are free to download and use 
's meme wherever you want (for non-commercial use only).
Please give credit and link back to the site.
Notes
- Spencer, Robert, August 10, 2026, Robert Spencer on X, @jihadwatchRS, Source
- EKO, "Jesus Grew Up Four Miles From a Roman City. Everyone Assumes He Never Left.", July 24, 2026, EKO LOVES YOU, Source
- Eugyppius, "Why AI slop is bad and people who post AI slop should feel bad", Eugyppius: A Plague Chronicle, July 24, 2026, Source
- John Paul Brammer, Kathleen Schmidt, Lincoln Michel, and Danya Issawi, "An author’s $2 million book deal was pulled over suspected AI use. Where does publishing go from here?", August 6, 2026, Unstacked, Source"
More quotes from this essay/roundtable:
"I’ve always thought the real issue with LLMs in publishing is quantity. They can pump out endless slop. This slop is already overwhelming things like Amazon’s Kindle Unlimited store. I also think we’ll see people turn to customizable LLM slop, paying OpenAI or whatever directly for stories. So I think literary publishers would be smart to make themselves a home for the people who want human-authored work.
Also… if you use AI in significant ways and are open about it, great! Readers can judge with their eyes and dollars. The problem is that very few such people want to admit what they are doing. That’s the root of all these scandals, really. It’s all the people hiding or lying about their use.
I think quantity is the threat. It can produce endless customizable work for very cheap. If AI capabilities improve as all the boosters claim, then why do you need a human there at all? If an LLM can produce a book that’s 95% as good as your average commercial thriller, how many people care about that last 5%? Some will. Many won’t.
Many, many people romanticize the idea of being an author. They think you can get rich from publishing books. What they don’t understand is that writing is a craft that people work at for years. AI undervalues that notion, and I don’t think non-writers quite understand that having an idea for a book and writing a book are two very different things.
And secondarily, I blame the AI authors who don’t disclose their use. We wouldn’t be having these problems if everyone were open about it, and the reason they aren’t open about it is because readers have so far been rejecting LLM books. So they lie or hide it to try and basically trick readers. I’m sure they have mental justifications for it, but it still is what it is." - Rosenthal, A.C., Source
- Eugyppius, "Why AI slop is bad and people who post AI slop should feel bad", Eugyppius: A Plague Chronicle, July 24, 2026, Source
- Here's the comment on the essay:
"@The Inmate,
First, thank you. You've recommended my work generously, publicly, and more than once, and you did it again inside a piece questioning whether I even wrote it. That takes a kind of honesty most people don't have, and I'm not going to repay it with a fight. So let me just answer you plainly, the way you asked. No worries.
You said you want a soul in every word. So do I. Let me tell you what my process actually is, and then you can decide.
I use AI the way I use spell-check. It's a tool, not an author. It might hand me a skeleton from a pile of bones, or help me articulate a thought I'm circling. But the pile of bones is mine and I go through every word of it. Every paragraph, every sentence. I add, I cut, I rewrite, and I rewrite again, until the thing on the page is me. If a line doesn't sound like me, I work it until it does or it goes in the bin. The tool never has the last word. I do. Always.
Now think about how any book you've ever loved actually gets made. An author writes it. Then editors reshape it. Proofreaders correct it. A publisher decides what stays. Typesetters and formatters lay it out. If it's translated, a translator rewrites every sentence into another language entirely. Half a dozen hands touch the words before they reach you, and we still call it the author's work. Because it is. The author had the vision and the final say, even on a translation they can't read a word of.
I compressed that entire apparatus into one tool. And unlike most published authors, no one overrides me. No editor kills my best line. No publisher softens my argument. No one has final say but me. By that measure, my work is more my own than most of the books on your shelf, not less.
On the detector, I'll say this once and leave it. A friend of mine has sons who build AI-detection software. Her own writing, every word of it her own, scores 95% AI. She laughs and takes it as a compliment, because what those tools actually flag is clean, competent, pattern-consistent prose. Good writers set them off, because of what the detectors measure, patterns. That's not my defense. My process above is my defense. It's just context for a number you're putting a lot of weight on.
And here's the part I'd ask you to sit with, as a friend, not a foe. Every image in your piece is AI-generated. You describe to a machine what you want to see, and it makes it. That's fine. It's a tool, and you used it well. But notice: you cannot get in there with a paintbrush. You can't touch a single line of that drawing. You describe, and you accept what comes back, even if you send it back fifty times. The image is still 100% AI. It's the machine's hand, start to finish. The only human mark on it is your watermark. HOWEVER, they are still YOUR images. If I was to use them you wouldn't say I used AI images, you would rightly say I used YOUR images.
Now, when it comes to my writing, I get in there with a keyboard for what the AI generates. Line by line, with my own hands, until the words are mine. In a way that is significantly different from the AI images. I don't just edit, I take a seed and grow it by hand.
I'm not arguing about which of us has more of himself in the work. I'll grant that you probably do, since you claim to use no editing software at all. But for those same reasons, I believe I must have more of myself in my own writing than most books do, whether they used AI or not. Because I did all of my own typefacing, editing, proofreading, publishing, and made my own book covers. I reworked every sentence by hand until it carried my own voice, and rewrote them many, many times.
I think this is the same standard you'd want applied to yourself. A tool touched it. But a mind shaped it, and a person stands behind it. That's true of your images, and it's true of my essays.
I stand behind every word. That's the only claim I've ever made, and it's the true one.
Still glad to be recommended by you. I stand behind my work the same way I have stood behind my work as a tradesman, whether I used a power tool or did the work by hand. Please let me know if i can review something of yours in return. I'd be happy to.
A. C. Rosenthal"
This one flagged as 51% AI and 49% human. The human part begins at "It's just context for a number..."
And there's that, "A friend of mine has sons who build AI-detection software..." LOL. Not just a son, but sons! Their software is so good she gets flagged 95% of the time—because she's such a good writer? That is just too convenient.
I've also noticed over my years on Substack that AIs often give long responses as a general rule. Trolls do the same thing, but I suspect many if not most of those are AIs. - This is the comment on my review of his book:
"The Inmate,
Setting everything else aside, I want to thank you for the review itself. You read the book the way I hoped someone would read it. You've taken my work seriously. Most people skim for the thesis and move on. You actually READ it.
You've got the central claim exactly right, that the God most people reject was never the God of scripture in the first place. They reject a caricature someone else drew for them, and then they spend years angry at a cartoon. If a reader walks away with only that, I've done my job.
But the part that stopped me in your reply was that you made the message personal. You left the religion at around 40 and came back in your 60s with your doubts still intact. That's the reader I was writing for. I have a person in my life who I wrote this book for, and I've not had the courage to share it with him yet. I'm scared that it would not be received well, and I love him like an uncle. Your review is most encouraging to me. He is also not the kind of person who wants their questions silenced. The fact that the Prodigal Son section landed for you, specifically, given that history, means more to me than any number in a stats dashboard. That parable is the hinge of the whole book for exactly the reason it moved you: the father runs while the son is still filthy, still rehearsing his apology, still expecting to be made a servant. The father doesn't wait for the son to clean up. That's the God people miss.
And I'll say something about coming back with doubts intact, since you were honest enough to raise it. I think that's the only honest way to come back. Faith that requires you to first kill every question isn't faith, it's a performance. The doubts don't disqualify you. They're evidence you're taking the thing seriously enough to wrestle with it. Jacob walked away from that fight with a limp and a blessing, in that order. I'd trust a man who limps over a man who claims he was never wounded.
You noticed the book doesn't end with an altar call, and you were right that it's deliberate. I'm not after a conversion moment with the organ playing. I'm after consideration. Come, let us reason together, as the prophet put it. Or more plainly: "Have you thought about this?" That's the whole posture. If someone closes the book and merely reconsiders what they thought they'd settled, that's enough. The rest isn't my work to do.
Thank you for reading it with that much care, and for being honest about your own road through it. That's the conversation the book was written to start. I don't have the words to express how encouraging this is. Thank you."
Very nice. This one is flagged as 100% AI. - BIG WILD MIND MACHINE, "The Man Who Occured[sic]," June 22, 2026, Source
- Rosenthal, A.C., "Skepticism Is Not Indifference", June 29, 2026, Source
- Nabu, "What the Gospel of John Was Actually Saying", Earth Mastery, July 23, 2026, Source
- Chaudhari, Deepshikha, "The Color That Stayed on My Hand", December 18.2025, Source
- Rosenthal, A.C., "The Most Falsifiable Holy Book in the World", June 10, 2026, Source
As censorship increases also consider using email and text messages to send links.
Leave a Reply