I’m sure you’ve noticed—it’s everywhere. The dreaded em dash is taking over every post, tweet, caption, and even some books! This essay is not meant to hate on the em dash, for it’s just an innocent punctuation mark in our punctuation toolkit. In fact, it’s a useful piece of punctuation—it can make our writing more casual and conversational, adding thoughts to previous ones without the serious stop of a period. It adds more drama, even more rhetorical flair, and a longer stop than the mini pause of the comma. Instead of being restricted to joining two independent clauses like a chunky semicolon, the em dash is flexible; it can connect an independent clause to almost anything—another independent clause, a dependent clause, a phrase, or even a single word.
So, what’s the problem? Why is it “dreaded”? The problem is that AI has taken it hostage. The em dash is the most favorite punctuation mark of our sly robot writing friends. I’ve spoken with ChatGPT about its obsession with the em dash; you may have too, but it told me that the em dash is its “punctuation soulmate.” Sigh. It favors it entirely over the “chunky” semicolon. It doesn’t hide its love for the em dash, forces it upon us in obscene quantities, and even continues to do so after being told not to (based on real experience). Every time I put any kind of writing into ChatGPT, it consistently uses as many em dashes as possible. But, if it’s a great punctuation mark, shouldn’t we use it more? Let them use em dashes, you say!
My issue is that we have used it to teach AI, and now AI is overusing it. We aren’t thinking about checking AI's punctuation work. I would argue that the range of the em dash is being exploited and overused by AI to the point that now, when I see it (multiple times) in any writing, I question if that writing was written by AI. If I see other signs, such as gushy yet vapid vocabulary or trite expressions, I usually know it’s AI. While I’m not a professional editor, as a writing tutor constantly reading student writing with an eye for AI, I have become increasingly aware of the nuances between human-generated writing and the voice of generative AI. Yes, AI is easy and saves time, but at what cost? Our voice.
The saddest part of the em dash hostage crisis is that we are giving over our voice to AI, our writer's voice. We cannot give over our voice to AI. Punctuation makes up syntax, and syntax forms our voice, and our voice communicates our ideas. If we let AI rule our punctuation toolkit, our voice will become very… one note. If it hasn’t already…I hate to see people who have incredible voices, vocabularies, and viewpoints outsourcing their writing, aka their voice, to AI. It’s spitting it back out with a combination of their ideas, but it sounds like it has been autotuned to sound like everyone else. It saddens me to see a generation of writers turning to AI not only for ideas but to construct their written voice. The product: everyone sounds the same. I don’t want to live in that world. I want to live in a world where we are brave enough to make mistakes in our writing, use punctuation creatively, and take the bold step to put forth our ideas in our own voice.
Before I get too sad about this reality, let’s get back to the em dash. It’s in need of saving, after all. The main problem is that it is being overused. I’m not against using here and there, maybe once in an Instagram or LinkedIn post, but not 3-5 times in every post you write. I’m not against using it when writing fiction or nonfiction, but maybe not 4-5 times on every page (yes, I’ve seen this). Just like words and phrases in our writing, we need to vary the kind of punctuation we use, too. What if ChatGPT was obsessed with the period? What if it only wrote short sentences? Then we would get fatigued. We would think there are too many stops. Keep the thought going. I would scream! Add a comma, use a semicolon, or an em dash for heaven’s sake. It’s the same with the em dash or any object, idea, or style we overuse—it’s fatiguing and develops associative meanings. The em dash is now associated with AI and ChatGPT—use at your own risk. Perhaps try your voice instead. You might be surprised to hear what you say.
Comments