Is ChatGPT Changing the Way We Write?

Is chat GPT killing the M dash

Generative AI is doing a lot of harm. Environmental, privacy-related, even potentially affecting the brain activity

of those who use it all too often. So the potential demise of a punctuation mark

may seem to pale in comparison, but that doesn't mean that it doesn't matter

In typography, an M is a measuring unit. So an M dash is literally a dash that is one M long

The punctuation mark has been used for centuries in English, shaping playwriting, prose and poetry

You can famously find it in the poems of Emily Dickinson, in Shakespeare, Laurence Stern, Virginia Woolf, you name it

The list is long and impressive. Today, the M-Dash is mostly used in writing to elaborate on thoughts, give examples, add supplementary information, explain or expand on an idea

However, lately, the excessive use of mdashes has been cited as one of the main giveaways

that a text may be written by ChatGPT. In fact, ChatGPT generates so many mdashes that some Gen Z users have been referring

to it as a ChatGPT hyphen. I rarely use ChatGPT, and when I do, it's never to generate text

And yet, I have caught myself editing mdashes out of my writing, swapping it for commas

or other punctuation marks, sometimes even completely changing the initial structure of my sentences

And I really have a problem with this. I believe that the way a person writes tells us a lot about their thought process

And the M in particular with its poses its hesitations and interjections mirrors real human speech to me And so I hate the idea that an AI platform that has been trained on human writing will ironically force people to change

the way they write. And it's not just about self-censoring. In a piece for The Guardian

journalist and fellow M-dash lover James Shackle writes about receiving an email from his editor

that called journalists to stop using M-dashes in their writing to avoid AI accusations

But why is OpenAI's platform using so many M-dashes in the first place

Here's what Sam Oatman said when asked about it. You know, we have this team that figures out

what the model's personality should be like and how it should behave. And a lot of users like M-dashes

So we had more M-dashes. And now I think we have too many M-dashes. But that's the answer

It was just like, users liked it. We put more in. Generative AI feeds on pre-existing texts written by humans

journalists, writers, bloggers, academics, anything it has access to. And so if humans use a certain punctuation mark a lot, the AI will too

Following considerable backlash and quite a lot of memes, in November 2025, Othman announced that OpenAI had fixed its M-problem

by adding a custom option to prompt ChatGPT not to use them

We're yet to see if the company will phase out its now-characteristic M-dash

If it does, in a few years' time, the ChatGPT hyphen may become a distant memory of the past

and we can all go back to using M-dashes without thinking of AI. Until then, please keep using the M-dash

There are other and more reliable ways to tell if your text was written by AI

and I would really hate to see the M-dash go