Ann Handley Is Not Giving Up the Em Dash

The MarketingProfs chief content officer talks about AI tells, disclosure, personal stories, and why she refuses to let a robot tell her how to write.

Ann Handley is not giving up the em dash.

She knows it has become one of the supposed signs of AI-generated writing. She also thinks that conversation has gotten a little ridiculous.

Handley, chief content officer at MarketingProfs and author of Everybody Writes, has spent much of her career helping marketers write with more clarity, voice, and humanity. Now, some of the choices that make writing feel personal — punctuation, rhythm, familiar turns of phrase — are being treated as evidence that the work may not be human at all.

In this companion Q&A to The Signal’s feature on marketing judgment, Handley talks about AI tells, disclosure, what writers should be careful not to hand over too quickly, and why the best response to generic writing is not to sand down your own voice.

This conversation has been edited and condensed for clarity.

The Signal: You recently wrote about the em dash, and I loved it because I am also an em dash loyalist – I love the drama. What do you make of this moment where punctuation, style, and voice are being treated as AI tells?

Ann Handley: I am so fatigued by this whole conversation.

You see it all the time on LinkedIn, especially. People say, “Here are all these AI tells. Watch for them.” The em dash is part of it. The word “quietly.” The short question followed by a rhetorical answer. Somebody called out “here’s a thing no one talks about.”

Come on, people.

None of those things are AI slop identifiers. Will they appear in AI writing? Sure. But that doesn’t mean they should be entered as evidence into some kind of public trial about AI use.

All of those things are also simply language. AI lifted all of that from us. Not the other way around.

That’s how we speak, and it’s how we write.

The Signal: So you are not giving up the em dash.

Handley: The em dash has become a hill I will die on, except it’s really just a hill I celebrate on because I refuse to give it up.

It’s so much a part of my voice. I just finished writing a new book, and I’ve been going through the copy edits. Girl, I use a lot of em dashes.

Do I overuse them? Maybe. But it’s the way I write. It’s also the way I speak. I tend to half-interrupt myself constantly. My brain is racing ahead, and I interrupt myself with an em dash.

I refuse to have a robot tell me how to write. I absolutely refuse.

The Signal: The AI-tell conversation feels connected to trust. People don’t always know what they’re looking at anymore, so they start scanning for clues. How do writers get through that without giving up the parts of their style that make the work feel like theirs?

Handley: I think it’s silly, what’s going on now.

You see it with LinkedIn’s AI slop button. You see it on Substack, where the machine will tell you whether your writing seems human or not. And then you hear about actual writers dropping typos into their writing as a desperate attempt to seem human.

I’m not playing that game.

I think it’s incumbent on writers to know when to choose the tool and when not to.

For me, disclosure is part of it. Be clear about how you use AI. Put it out there. Disclosure is one way to counter a lot of this ridiculousness.

The second thing is being clear about when it helps you and when it doesn’t.

Reaching for the tool is the impulse I’m trying to interrupt right now. It comes down to disclosure and being intentional about your own choices as a creative, a writer, and a marketer.

The Signal: You said the wrong question is whether AI can do something. Why is that the wrong place to start?

Handley: Because yes, probably AI can do it.

The better question is: What do we lose when we ask AI to do this for us?

That’s the question I think all of us in marketing need to ask ourselves. Is there an easier and faster way? Yes. But is there a cost to it? Also yes.

Sometimes you might choose that shortcut. Other times, maybe the cost is worth it.

But I’m worried about losing our capacity to do creative written work if we constantly reach for AI first because it’s faster and easier.

The Signal: You also used this great phrase, the ‘golden retriever effect,’ to describe how AI can feel in the writing process. What did you mean by that?

Handley: It feels very fun and easy to use.

The AI will just keep dropping tennis balls at your feet: “How about now? How does this sound? What do you think about this?”

And I don’t want that in my process.

When you use AI for a first draft, it’s like living inside a planned community as opposed to pitching your tent in the wilderness.

When you live in the wilderness, it’s uncomfortable. Your ankles get scratched. You’re sweaty. It’s hot. It’s not great. But that struggle is the process – that is the process of creating.

I don’t think it’s too high-minded to say the struggle is important for us, even if we’re just writing a blog post. It’s important to us as people to figure out what we want to say because writing is thinking.

The planned community is what AI will give you. It’ll say, “Here are the talking points to include. Here’s the general layout.” It’s fine. It’s pleasing.

But it doesn’t put enough of me into the work. It doesn’t make me care about the work, and it doesn’t create the connection to the work that I need.

The Signal: That makes me think about how much writing is about finding a point of view, not just producing copy. What parts of the process should writers be careful to protect?

Handley: I’m always leery about giving a prescription across the board, but I think marketers should build in human moments when the stakes are high.

For me, the stakes are highest in the first draft.

AI is fine to use for some aspects of research, for improving a piece after it’s down, for helping you optimize it, for writing CTAs or social posts. Those things can be useful.

What you need to protect are those human checkpoints, those moments when the stakes are high for you as a person.

I don’t think there’s a clean prescription for every team. But what I always tell my team is: You can say no. You can say no about using AI.

The Signal: What does human writing look like now, especially in B2B, where so much content can already feel generic?

Handley: I keep coming back to the fundamentals of great content.

In the second edition of Everybody Writes, the formula is utility times inspiration times empathy equals great content. There should be elements of those things in everything you write.

Utility means understanding what problem this is solving for the customer. Inspiration matters because AI has raised the bar – there needs to be more of us in what we’re producing.

One of the ways I do that is to tell stories from my point of view. Not to indulge myself, but to make me, my brand, and my company more human to the people reading me.

One piece of feedback I get all the time about my newsletter is that people feel like I’m writing just to them. That’s honed over years of judgment, but it’s also because that’s how I approach it. I try to think of a story or something that happened to me that can connect with the audience and reflect their problems, ideas, and lives as well.

The fundamentals still exist in the age of AI. The bar is just higher.

The Signal: What is one small writing choice that can make a piece feel more alive? And don’t say em dashes.

Handley: Em dashes.

Also, a story. A personal story that could only have happened to you, but also has relevance for everyone else.

The best stories are specific enough to be personal, but also universal. They have relevance that other people can relate to.

That’s something AI can never truly offer. It can make something up, but it always feels weird. As a reader, you can tell when something is fake like that.

The more specific and more personal you can make your content and your writing, the more it becomes your biggest trust signal.