The Em Dash Divide: One Colleague’s Obsession, and My Quiet Resistance
- Claire Roper
- 1 day ago
- 3 min read
We all have that one colleague—the one who insists an em dash is the most elegant piece of punctuation ever invented. For them, commas are cluttered, parentheses are passive, and colons? Don’t even bother. The em dash, they say, is the perfect bridge between thoughts. It’s dramatic. It’s clean. It’s versatile. But lately, a curious pattern started to emerge. Not just in their writing, but everywhere.
Marketing emails. Press releases. Blog posts. Even social media captions. That same stylistic signature—one long line, slicing through sentences. The em dash was everywhere...

When Should an Em Dash Actually Be Used?
Let’s pause for a reality check. The em dash isn’t just a stylish flair—it has legitimate grammatical uses. Here’s when it’s appropriate:
To set off parenthetical information Similar to commas or parentheses, em dashes can insert additional information without interrupting the sentence flow too much.Example: My colleague—a true em dash enthusiast—refused to use any other punctuation.
To indicate a sudden break or interruption Great for dialogue or when a thought is cut off mid-sentence.Example: “But I thought you said—” “I changed my mind.”
For emphasis or dramatic effect at the end of a sentence Especially when you're adding a surprise twist or punchy phrase.Example: There’s only one punctuation mark that gets this much attention—the em dash.
Enter AI, Stage Left (With a Typing Speed of 1,000 Words a Minute)
It turns out, the latest wave of AI-generated content leans heavily on em dashes. Platforms like ChatGPT, Jasper, and Copy.ai seem to have a soft spot for them. Why? Because em dashes give writing rhythm. They add a conversational tone. They signal a shift in thought without the clunkiness of more formal structures. Basically, they mimic how we talk—off the cuff, trailing into new ideas, looping back. Which makes AI writing sound incredibly human. So human, in fact, that it’s getting harder and harder to tell the difference.
Critics dubbed the em dash a potential “AI tell,” but expert writers caution that this association is superficial: any patterns arise from training data, not an inherent AI preference - Washington Post
That assumption, “I know this was written by AI because of all the em dashes,” has become a kind of shorthand for spotting machine-generated content. But honestly? It's a bit lazy—and often wrong.
Let’s Clear This Up: Em Dashes Don’t Always Mean AI
There’s a growing belief online that if a piece of writing uses too many em dashes, it must be AI-generated. And while it’s true that many AI tools lean into this stylistic choice, let’s not forget something important:
Humans were overusing em dashes long before the bots showed up.
Some writers genuinely love them (looking at you, dear former colleague). Others use them to avoid choosing between a comma or semicolon. Some lean on them for rhythm, drama, or flow. And in casual writing—like blogs and social media—they can make sentences feel breezy and conversational.
So no, an em dash isn’t a dead giveaway that AI wrote something.
Articles Explaining the Use of an Em Dash
The Punctuation Guide – "Em Dash" www.thepunctuationguide.com/em-dash.html
Grammarly – "When to Use an Em Dash" www.grammarly.com/blog/punctuation-capitalization/why-you-should-love-the-em-dash
Merriam‑Webster – "How to Use Em Dashes, En Dashes, and Hyphens" www.merriam-webster.com/words-at-play/em-dash-en-dash-hyphen-usage
What It Means for Writers (and My Em Dash–Loving Colleague)
Working with someone who naturally writes like an AI is a reminder of just how close we’re getting to that tipping point. Their style—once unique—is now echoed by every bot and algorithm out there.
AI is here to stay. Em dashes, too. But the question isn't whether a machine can write like us—it’s whether we’re still writing like ourselves. So the next time you read a blog post or email that sounds almost too polished, take a closer look. And if you spot three em dashes in the same paragraph, it doesn't mean it's AI, it could me, it could be you or even my former colleague’s latest article...
Comentarios