FIR #481: The Em Dash Panic — AI, Writing, and Misguided Assumptions
Description
In this short midweek episode, Neville and Shel dive into one of the hottest debates in communication today: what happens to tone and authenticity when artificial intelligence steps into the writing process? From the surprisingly heated arguments over the humble em-dash to fresh research on AI’s “stylometric fingerprints,” we explore whether polished AI-assisted prose risks losing the human voice that builds trust. Along the way, we look at how publishers like Business Insider are normalizing AI for first drafts, how communicators are redefining authenticity, and how Shel used AI to turn years of blog posts into a forthcoming book.
Links from this episode:
- Human-AI Collaboration or Academic Misconduct? Measuring AI Use in Student Writing Through Stylometric Evidence
- Distinguishing AI-Generated and Human-Written Text Through Psycholinguistic Analysis
- Some people think AI writing has a tell — the em dash. Writers disagree.
- AI is breaking my heart: Why authentic writing matters more than polished words
- The Em-Dash Responds to the AI Allegations
- Business Insider reportedly tells journalists they can use AI to draft stories
The next monthly, long-form episode of FIR will drop on Monday, September 29.
We host a Communicators Zoom Chat most Thursdays at 1 p.m. ET. To obtain the credentials needed to participate, contact Shel or Neville directly, request them in our Facebook group, or email fircomments@gmail.com.
Special thanks to Jay Moonah for the opening and closing music.
You can find the stories from which Shel’s FIR content is selected at Shel’s Link Blog. Shel has started a metaverse-focused Flipboard magazine. You can catch up with both co-hosts on [Neville’s blog](https://www.nevillehobson.io/) and [Shel’s blog](https://holtz.com/blog/).
Disclaimer: The opinions expressed in this podcast are Shel’s and Neville’s and do not reflect the views of their employers and/or clients.
Raw Transcript
Shel Holtz (00:01 )
Hi everybody and welcome to episode number 481 of Four Immediate Release. I’m Shel Holtz.
@nevillehobson (00:08 )
And I’m Neville Hobson. In this episode of For Immediate Release, we’re going to explore the question of tone and authenticity when artificial intelligence becomes part of the writing process. That seems to be a bit of a hot topic these days from what I see online. AI tools don’t just generate text. They also polish, rewrite, and shift tone to make communication sound warmer, more professional, or more concise. But what happens to authentic voice when AI smooths the edges?
Do we risk losing individuality, nuance and trust if everything starts to sound the same? We’ll talk about that right after this.
It’s a debate playing out among communicators. This year, the humble M-dash has become a flashpoint. Some insist that overusing M-dashes is a dead giveaway of AI altered text. Others push back saying that’s nonsense and unfairly stigmatizes a perfectly good mark of punctuation. Washington Post ran a feature in April with the headline, Some people think AI writing has a tell. The M-dash writers disagree. Then in August, Brian Phillips wrote a lyrical defense in the ringer.
pleading, stop AI shaming our precious kindly M-Dashes, please. And McSweeney’s even joined him as satire, publishing the M-Dash response to the AI allegations written from the dashes own point of view. That is really, really very amusing, worth a read. The fact that such debates exists highlights how sensitive people are to the signals of authenticity in writing. Fresh research in 2025 suggests this is more than speculation.
Some recent studies show that AI leaves stylometric fingerprints in writing that can be detected, raising questions about authorship and voice. A stylometric fingerprint is the unique combination of statistical linguistic features within a piece of text that identifies its author much like a human fingerprint. AI can make writing clearer and more polished but risks homogenizing style and raising ethical questions. Beyond academia, commentators argue that polished words without voice
risk-leaving communication hollow. And while researchers are busy analyzing stylometry and psycholinguistics, communicators are having a very different kind of debate about punctuation. So while academics study the fingerprints AI leaves on writing, the popular imagination has latched onto something much simpler, the punctuation choices we make. The M-debate may be tongue in cheek, but it speaks to a serious point.
How sensitive we’ve become to the signals of authenticity in text right down to a single line on the page. For communicators, the challenge is not whether to use AI, that ship has sailed, but how to preserve authenticity when tone shifting tools are in the mix. The call to action is to define what authenticity means in your context, decide which writing tasks AI should support, and ensure human voice and accountability remain front and center.
In the end, and authenticity aren’t about perfectly polished words. They’re about whether people believe there’s a human voice and accountability behind the message. Your thoughts, Shale?
Shel Holtz (03:19 )
I have a lot of thoughts on this, ⁓ beginning with, well, at least it’s not the Oxford comma, because that’s a source of debate without artificial intelligence. The dash has found its way into AI outputs because of the AI inputs. The training sets include this massive corpus of writing that has been scraped from the web.
@nevillehobson (03:28 )
Don’t start on that.
Shel Holtz (03:49 )
that includes dashes. This is what it learned from. It saw a lot of ⁓ dashes used in writing, and that’s a pattern that it recognizes. It recognizes where they’re used and implements it in the outputs it creates. It did not think to itself, you know, I haven’t seen many dashes in my training set. That’s a shame. I’m going to start using more of those. That’s absurd.
When I was in college, I had a part-time job setting type. Yeah, I’m old enough that I actually set type. And I remember learning when to use the ⁓ dash based on the copy that I was transcribing into typeset. And there were a lot of them, even back then, when most people were still working on typewriters. So I think this notion that it’s a tell is ridiculous.
It’s, as you quoted somebody saying, a perfectly serviceable bit of punctuation. But let’s go beyond this notion of punctuation. mean, leave it to communicators where we have some really weighty issues to deal with that we’re going to spend most of our time talking about punctuation. I think this is one of the reasons leaders don’t take communicators very seriously. They’re thinking about business decisions and we’re thinking about
⁓ letting and kerning and things like that. We really need to get more focused on business outcomes and how communication and the use of AI contributes to that. This is what the Business Insider has done. Business Insider, I don’t know if you saw this, it’s a very recent announcement, have ⁓ officially given permission to their journalists to use AI to write the first drafts of their articles. They were already…
@nevillehobson (05:44 )
Yeah.
Shel Holtz (05:47 )
allowed to use AI for research, but not for any of the drafts they produced. Now they can produce the first draft. Now, why was that decision made? I’m not privy to what was going on in the mind of the president of Business Insider who made this decision and communicated it through internal memo to her staff. But I have to believe a couple of things are in play. First of all, journalism is in financial difficulties and you need
fewer people to crank out more stories. And if you can get it done faster by having AI generate a first draft, you then go in and fact check and clean up and apply your own writing to, I’ve done this, I’ve done first drafts in AI. And by the time I’m done rewriting, it’s a completely different piece. Still took me about an hour and a half less than it would have if I had had to sit there and write the first draft. I don’t do that for every kind of
article or other material that I need to produce, but on some things it just makes sense and it makes life easier. But the other reason I think Business Insider decided to go down this road is because AI is getting better at producing these kinds of drafts and it’s going to continue to get better. In the world of business, and I’m sure you’ve had this experience Neville, I know I have, is outside of the world of
communication when you are dealing with people in other parts of the business. And I don’t mean this in any sort of pejorative way. This is not an insult. These people are brilliant when it comes to the areas of specialization that are the focus of their jobs. But they can’t write their way