I have no interest in reading AI-generated content. AI-generated content is artificial, and because it is artificial, it lacks soul. That which lacks soul cannot speak to souls.
That’s the AI stuff you see.
You can’t see the good stuff.
See: survivorship bias.
Not according to this statement.
The en-dash (option-shift-hyphrn) is normally reserved for numeric ranges, like page numbers or dates, or scores, or martial conflicts between two factions or nations, or contrasts or occasionally, under complicated circumstances, hypenated phrases including prefixes and compound nouns (Chicago Manual has a helpful list).
I suspect but do not know that part of the LLM use of em-dashes is becausev they were partly trained on pre 1920 writing that’s open source, when em-dashes were much more common.
I’m as susceptible to being fooled as anyone; nevertheless, I have no interest in reading probabilistically compiled prose. In my view, it has no inherent value.
Are you talking about the stuff where someone say asks Claude to “write me an 800 word article about religious schools and bottlenecks” then publishes it?
Or are you totally black and white about this?
I wasn’t being entirely serious in the OP! It’s all a question of house style, and ours in the UK are more likely than not to use en-dashes.
In general it’s a national convention thing. American publishing houses mostly use no-space+em-dash, and a few British houses—such as the OUP—still do.
But you’re (much, I think) more likely to see space+en-dash+space in British published work, including newspapers. Wikipedia reckons that others, such as Canada, also favour the en-dash.
We all still use en-dashes between dates of course.
E.g. Go to this page ( UK position on Falklands will not change, No 10 says after leaked Pentagon memo | Falkland Islands | The Guardian) and search for an en-dash and you’ll find 5 en-dashes with spaces and no em-dashes. That’s not an error: it’s their house style, which they publish, and it’s entirely normal for a UK publication.
I have observed that AIGC suspicion is usually evoked by generic talk about generic points.
We are tempted to use threes when we want to convince but don’t have a straightforwardly compelling argument.
We dose an essay with ornate wording when we believe form is as important as content.
We write This is not X but Y when we fear of being misunderstood or scorned. Negation is cheap. Accurate description of what you’re talking about is cognitively expensive.
The LLM is a cheap writer who makes halfhearted attempts at convincing you. If you do not want to be mistaken for an AI, don’t be that kind of writer. Concision is the best rhetoric in this Age of Too Many Words.
I apologize in advance for my long response, but you asked. 
Good question, and not a black-and-white one for me. The “write me an 800-word article about X” workflow is exactly what I had in mind, and yes, that is what I find hollow. I would also say that having AI generate an initial draft the author subsequently edits, even heavily, is problematic. I am more comfortable writing or dictating the initial draft and then editing it than having AI generate the draft I edit.
I draw the line at ghostwriting, not at AI assistance generally. I use AI quite a bit as an editor, sounding board, and research aid. The distinction I try to hold is between AI doing the thinking and writing for me versus AI helping me sharpen what I have already thought and written.
I wrote about this at some length a while back. The relevant portion of that article is below. It captures my reasoning better than I can summarize it here.
Excerpt from “Ghost Writers in the Sky: Navigating AI’s Role in Authorship”
Using AI as a ghostwriter should scare you.
While AI has tremendous potential to be helpful, it also comes with great peril. Using AI as a ghostwriter may bring “efficiency and convenience,” but it also comes with the potential for significant loss.
AI can mimic tone and style, but it lacks the lived experiences that shape and deepen your writing. It assembles words based on probability, not reality, resulting in content that feels hollow. Worse still, relying on AI as a ghostwriter fosters impostor syndrome. You become a specter, an echo, a shadow of yourself. Writing is a craft that requires effort and practice. If you let AI take over, your ability to engage deeply with language, ideas, and your readers will atrophy.
AI ghostwriting raises serious ethical concerns. Does AI deserve credit for AI-generated work? Is it deceptive to claim AI-assisted writing as entirely one’s own? Does AI dilute the value of human authorship?
I believe the answer to all three questions is yes. Claiming AI-generated material as your own is disingenuous and fundamentally dishonest. Using AI as a ghostwriter and passing off its work as your own crosses the line of intellectual integrity and diminishes you as a creator. When AI generates content in place of your own, it stifles creativity and short-circuits original thinking. Your unique voice is replaced with something artificial, generic, and formulaic.
Writing is not merely about producing words. It is about shaping ideas, wrestling with language, and refining thought. Writing is a means of thinking, reflecting, and growing. It sharpens reasoning and deepens self-awareness. As Donald Murray puts it:
I write to say I am, discover who I am, create life, understand my life, slay my dragons, exercise my craft, lose myself in my work, for revenge, to share, to testify, to avoid boredom, and to celebrate.[1]
When AI does the work for you, you forfeit these benefits. Surrendering the writing process to AI weakens both the writer and the writer’s work. Do not lose yourself by letting AI write for you. Do not waste your mind by outsourcing the struggle and reward of writing.
Readers engage not only with the words themselves but also with the person behind them. Artificial intelligence cannot replicate the depth of human experience or the full range of emotions that shape our lives and give rise to our words: grief and joy, love and loss, hope and despair. Authentic writing carries the imprint of the author’s journey, reflecting the struggles, convictions, and insights that make it real. An AI ghostwriter, no matter how advanced, can never forge a true connection with the living.
Most of us want to be genuine. We want to be people of integrity and honesty. We do not want to fall victim to impostor syndrome or to pretend to be something we are not. To the extent that this is true, using AI to do one’s writing will, unless one’s conscience is altogether numbed, create an undercurrent of guilt arising from the realization that what one has put out in the world is not authentic but artificial. Do not do that to yourself or to others.
Although I do not believe AI should be used as a ghostwriter, there are many ways to use AI with integrity, authenticity, and efficiency. Used properly, AI can make you a better writer, not by writing for you, but by assisting you.
Here are a few ways to use AI to enhance, not replace, your writing.
Authors have used editors for centuries to improve their writing. AI can be a powerful, free, always-available super-editor. It can check spelling and grammar, flag redundancies and clichés, suggest rephrasings, and more. This is no different than having an editor review one’s writing. The author may or may not accept all suggestions, but a good editor will make one’s writing better.
AI can help generate ideas we might not have considered on our own. For example, I asked an AI client, “Give me examples of how to use AI for ideation. Are there best practices?” It returned insights such as brainstorming discussion questions, structuring content, exploring different perspectives, and overcoming creative blocks.
AI can serve as a debate partner, challenging assumptions and refining one’s thinking. To test this process, I had AI argue against a position I held strongly. Exercises like this help evaluate the pros and cons of an issue, assess new policies, and test the strength of an argument in writing. AI-generated content should never be accepted uncritically, but engaging in a simulated debate can sharpen your thinking, stimulate new ideas, and shape how you write and communicate.
AI offers many other ways to support writers. The key is to use it as a tool for enhancement, not a substitute.
That said, I am still learning and assessing my use of AI for creative work. It can be a great tool, but like most things in life, a good thing can be misused. I am trying hard not to misuse it.
I want what I put out into the world, warts and all, to be mine, not the output of a sophisticated machine.
Murray, D. M. (1990). Shoptalk: Learning to write with writers. Boynton/Cook. ↩︎
It’s a bit ironic that the excerpt is clearly AI-generated, from head to toe.
I agree with much of what you wrote, but I do not agree that “we are tempted to use threes when we want to convince but do not have a straightforwardly compelling argument.” I have read numerous articles, which I will not take time to find, about the value of threes in writing and other creative endeavors. I have used threes for as long as I can remember. I hope I do so as an effective writing technique, not as a reflection of poor reasoning.
It is not. I wrote it and used AI to provide editorial suggestions. Why do you assume it is AI generated?
If they are actually so valuable, why are they seen as a hallmark of AIGC by so many tech-literate people with a decent education? Something doesn’t add up.
I believe “unnecessary” threes, even before the advent of AI, was the hallmark of modern scholasticism in professional writing. YMMV.
One has to remember that AI is trained on human writing, which often use threes. On the other hand, perhaps I was poorly trained but I don’t call any of my teachers from elementary school to grad school making an issue of the use of threes. Perhaps I had bad instructors?
It matched every one of my points above. Especially the sixth:
Why would this guy post such a long comment if he himself is aware that it’s unnecessarily long?
It’s unfortunate that my diagnosis was wrong, but I won’t apologize. If I give everyone the benefit of doubt then nothing is AI-written.
That’s because scholasticism is actually encouraged in schools, hence the name 
It’s not your fault. We are all victims of the disrepair of public education, which started long before school shootings and mental health awareness campaigns.
I don’t quite understand, but I’ll leave it there. If it is helpful evidence that I wrote the article, warts and all, here is the introduction that I excluded in my post due to its religious focus not being relevant to the question I was asked.
Johnny Cash’s haunting song Ghost Riders in the Sky has always resonated with me as a powerful metaphor and cautionary tale about the dire consequences of sin and the urgent need for repentance. You can listen to the song on YouTube and most music streaming services. Take a few minutes to listen carefully to the words.
Given that later in his life Cash professed to be a Christian, it’s likely this was the message he intended to convey. Mark Powell, in his article “The Rise, Fall, and Redemption of Johnny Cash,” recounts how Cash, despite growing up in a Christian home, abandoned his faith after achieving fame and fortune. His life spiraled into darkness, marked by promiscuity, drugs, and the end of his marriage. However, Powell and those who knew Cash best testify that the Lord graciously drew him back to Himself. According to Powell, from that point on, Cash devoted the rest of his life to proclaiming Christ through his music, films, and public speaking.
Cash’s rendition of Ghost Riders in the Sky became a massive hit, peaking at number two on the Billboard Hot Country Singles chart. Released on April 1, 1979, the song tells the gripping tale of ghostly cowboys doomed to chase an elusive herd across stormy skies. It serves as a stark warning to a wayward cowboy: if he doesn’t change his ways, he will join the damned riders, forever condemned to “catch the Devil’s herd across these endless skies.”
While the song’s primary message is spiritual, I am using it as a metaphor for the concept of “ghost writers in the sky.”
The song’s imagery evokes the influence of unseen forces shaping the world, akin to ghostwriters who work in the background, crafting words, stories, and messages that others claim as their own. Much like the spectral cowboys riding across stormy skies in Cash’s song, AI exists invisibly in the digital realm. It operates in the “sky” of the cloud, generating content, ideas, and even entire articles and books—just like a ghostwriter whose hand remains unseen.
The use of threes (Tricolon) is a rhetorical technique that’s been successfully applied for at least 2000 years. It can be used to bolster inadequate arguments, but then so can every other rhetorical device, including antithesis (“It’s not X; it’s Y”).
But the use of the technique doesn’t make the writing bad: the best writers of English you have ever read use these techniques all the time. Some obvious examples from speeches that have been universally recognised as exceptionally convincing.
In (correctly) pointing out that AI overuses the techniques, there’s a danger people overcorrect by not using them at all, which is just as likely to lead to a flattened, less effective style.
Too much concision is just as likely to lead to bad writing as verbosity: it can disrupt the rhythm and flow of a text, and therefore impede meaning. The secret is ‘the right number of the right words in the right order’…
They lived in an era of too few words. So they—the elite few who could write—wrote more, to the great delight of their readers.
The English language changes with time. Whether you like AI or not, you will likely agree that writing as a subject today is quite different from what it was before 2022. Just look at how college students, the people who will become the next generation of writers, scramble to get their course work through AI detectors.
Given the above discussion, this is a timely article, which I discovered serendipitously as I was reading Apple News.
“Horwitz is not unaware of the contradiction. “I want to live in a world where people still can distinguish AI writing from human writing, which is why it’s so ironic that I built this thing,” he says.
And still, whether people online treat the project as a tool—which Horwitz is charging $4.99 a month to use after a 3-email free trial—or just as a quick joke, its fast virality underscores a larger conversation. In a world fatigued and frustrated with AI, a human touch is valuable. As one user on X puts it: “Stop being ashamed of typos, embrace them, It’s one of the last things we have to ourselves.””
On the contrary, the best writers you read today use rhetorical devices all the time. Go and find the most convincing speech or article you have read this year and see how many standard rhetorical devices are used.
That doesn’t mean that any particular piece will use all the devices, or even a particular one; just that good writers use the tools of good writing, and they most certainly include the standard rhetorical devices that have worked for 2000 years, across all Western cultures.
BTW, I quite like the suggestion that JFK’s speech from 1961 represents the ‘elite few’ who were able to write in a period of few words…
Well, no, I don’t agree. You seem to be suggesting that everybody’s perception of what is good writing will change because today’s students are having to adapt the way they write in order not to be marked as having used AI. Is that it?
Do you think students are being taught to write deliberately badly just to pass this test? Wouldn’t that be a shocking dereliction of duty by university authorities?