The Strange Case of the Em Dash in the Age of AI Anxiety

I recently found out about quite a ridiculous thing going on in the writing community, and especially on popular platforms for writers. Yes, it’s about writers deliberately not using em dashes, that is, “—” or “ — ” depending on the style variation (US/UK/Formal/Loose), to prove that their work is not AI-generated. For a writer who’s been around these communities and, thankfully, escaped to his own platform and audience, it’s funny on so many different levels.

The first and foremost is that no actual reader who found an article or essay online cares or even notices the difference between the formatting symbols, because people don’t read — they scan. That might be hard to swallow for someone who loves to write, but that’s a simple truth about online publishing.

The illusion of the attentive reader

Anyone familiar with tools like Google Analytics understands what I’m writing about, because it’s measurable as a metric called “average engagement time per active user.” And you know you’re lucky if your “5-minute read” (based on 200–250 wpm) article exceeds 60 seconds of average engagement. One large study (based on billions of visits) shows that the majority (55%) spend less than 15 seconds on a page. Meanwhile, eye-tracking tools keep showing that the majority of attention lands on the headings and subheadings, the first few lines, and then quick scrolling for highlights, which forms the famous F-shaped pattern.

Think of it even from your perspective, my dear reader. How often do you actually end up in a situation, “Oh, there’s something on the web. Let me devote 5 minutes of my full attention to read it slowly and carefully.” Who the heck does that? Don’t we all listen to podcasts on 1.5–2x speed while doing other tasks? Don’t we live in an era of rush, multitasking, screaming headings, and attention economy? Do I like or support it? No, but does it change anything? Denial won’t change the reality of how split human attention is.

General reading, like actual books, might be a different case, because people treat it as a digital detox, but online reading has its own clear pattern. People search for some information and, in most cases, scan only headings and subheadings, or highlighted sections in the body text, to confirm what they’ve already found in the title and leave for good. That’s it. And here you are, deleting em dashes from your work to prove that your content is not AI-generated, like anyone even notices them. It’s ridiculous.

See, in school or even in college, it might be different and maybe even useful for kids to learn how to convey their thoughts in the form of essays without external tools — just like it’s useful to learn basic calculations without a calculator. But in the real world, after school, there’s no point in that.

Inside the echo chamber of writers

The nonsense with “non-AI writing” is mostly visible on popular platforms for writers, where the writers mimic isolated island tribes that refused to leave their shores as the rest of the world sailed on.

They live in their little communities and write in an echo chamber of writers, because there are very few actual readers on these platforms — I haven’t met one myself. However, there are lots of writers who write and engage with other writers to boost their engagement, trying to make a career or at least an extra buck from an old craft and stressing whether others notice their little AI touches. While no actual reader even pays attention to such things. What a tragedy that is.

How do I know? Because I’ve been there. Both as a writer and as an editor. And even before the AI era, they were stressing about the grammar, formatting, or using tools like Grammarly, talking about who’s real and who’s not — that is, who’s writing without extra tools. Also, they created a kind of power structure based on the indicators, using various built-in systems to exercise their power, saying what’s good or allowed and what’s not. All this time, it was merely an ego game of writers in a declining profession — or a last desperate power-grab before the inevitable extinction.

And all this time, no one, even for a second, thought about the readers, because if they did, they would quickly find out that no one cares and wants their perfection or complexity, but quite the opposite. Just like now, they don’t care how to use AI tools to create better things — i.e., how to convey complex ideas more quickly and more sharply, to make their writing more competitive — but delete the dashes just to prove they’re pure writers, while no one cares about it.

When AI breaks the old hierarchy

Readers want solutions, answers, ideas, something to relate to, or just emotions, and they follow individual creators who curate them. And AI is just another tool to help creators create their things. For example, for me, a non-native English speaker living in a non-English-speaking country, it’s a phenomenal tool to ensure that my essays don’t hurt your eyes, dear native speaker. I’ve done quite well on this for years, but now it’s even better. It’s also speeding up my research and offers me lots of feedback as an editorial assistant.

How in the world can this be bad? Well, for the “elite” writing communities, it’s an obvious threat — not because AI writes with soul, but because it writes well enough. Suddenly, more people can express their ideas clearly and beautifully, without asking the gatekeepers for permission. The monopoly over words they held for so long is dissolving, and that’s the real reason for their panic. This is why they invent new rituals of purity — what’s “allowed” and what’s not — to protect a fading hierarchy to their benefit.

But the whole idea of AI as an omnipotent tool is totally wrong. Yes, it produces content faster and often better than most writers I have encountered over the years — me included. It’s just better in the mechanical sense, just like a calculator is better at calculating or a knife is better at cutting than our nails. That’s what tools do. But what we do with the calculator or with the knife — especially the knife — is on our end, right? We can use it to do helpful things, to build, to create, or to hurt ourselves. It's like in the old Chinese saying: “When the wrong man uses the right means, the right means work in the wrong way.”

The real problem isn’t that AI creates — it’s that creation without understanding quickly collapses. You can play this trick once, but if you don’t know who you are or haven’t developed a voice and style, it will be some random AI slop each time. In other words, you can make a debut on the writing scene with something awesome, but then it will be inconsistent and confusing for readers, because the hand that holds the tool lacks judgment.

What still makes writing human

See, for me and the em dashes, I will be brutally honest — I didn’t even know what this cute symbol is called until recently, because I’m not educated in writing, especially US/UK style. And before discovering “ — ”, I mistakenly used “ - ” for the same purpose for years, writing in the local market as a full-time copywriter or creative writer. And you know what? No one ever pointed out that mistake.

Then one time in some text editor, I clicked “-” two times, and it automatically made this “longer and more awesome-looking line,” so I used it ever after — way before AI was there. I did that only because it looked cooler stylistically. But even as an editor, I never really cared about the symbol and saw it as a different form of stylistic expression. I never pointed it out as a mistake because I have always looked at writing as a form of expression, even a form of art.

Sure, I can be ignorant. But if a professional copywriter, writer, and editor who wrote and reviewed hundreds — probably even thousands — of articles/essays, including pieces for the local market, didn’t care about it for years, how can you expect that a normal reader will see the critical difference between using “-” or “–” or “—” (or with spaces around) and judge you for that? Isn’t it just a way to prolong the sentence with some line? Isn’t it just that — just a line?

I honestly feel sorry for all writers stuck in an echo chamber of writers that has made them think it really has any meaning for actual readers, whether they used the symbol correctly or created their piece with AI assistance. They add drama to their writing process, whereas most readers give them 15–30 seconds of attention and leave. And no, it’s not because they saw em dashes in the content, marked it as AI, and refused to read it, but because they found what they were looking for and simply didn’t care for more.

But the few readers who read your work from top to bottom are happy that you have adapted to the new tools and can produce more of what they’re looking for faster, with fewer mistakes and more clarity — just like people interested in reading science were once thrilled about the calculators and computers their favorite scientists used to understand the living whole a little more.