The Em Dash Isn’t Dead. AI Just Ruined Its Reputation.

I saw a post recently from a copywriter defending her use of em dashes. She’d been using them since she started writing essays at school in 1993. Writers had been using them for centuries before computers existed, let alone ChatGPT, so using an em dash certainly didn’t mean something had been written by AI.

She’s right, of course.

ChatGPT didn’t invent the em dash. Writers didn’t suddenly discover punctuation when generative AI arrived. And seeing an em dash in a LinkedIn post, website page or article is absolutely not proof that somebody asked AI to write it.

But I think there’s another side to it.

AI does love an em dash.

And after reading enough AI-generated content, you start recognising the patterns.

It’s not really about the em dash

The problem isn’t one particular punctuation mark. It’s what happens when all the familiar AI writing habits start appearing together.

The beautifully balanced sentences. The perfectly organised paragraphs. The repeated groups of three. The neat little contrasts. The endless variations of “it’s not about this, it’s about that”.

And, yes, quite a lot of em dashes.

There’s nothing technically wrong with any of it. In fact, that’s partly the problem.

AI is very good at producing writing that is grammatically correct, logically structured and easy to read. Give it a sensible brief and it can produce perfectly competent copy in seconds.

When thousands of businesses are using similar tools, trained on similar information and asking for similar things, you inevitably start seeing content that feels familiar.

Different business. Different website. Different subject.

Same voice.

AI has made us suspicious of good writing

I think this is one of the stranger consequences of generative AI.

It hasn’t just changed how businesses create content. It has changed how we read content.

Certain phrases, sentence structures and punctuation choices have become associated with AI simply because we’ve seen them used so frequently in AI-generated writing.

That doesn’t mean we should deliberately start writing badly to prove there was a human involved.

I’m certainly not going through articles removing every em dash or adding grammatical mistakes for authenticity.

But I do think businesses need to pay much more attention to whether their content actually sounds like them.

Using AI isn’t the problem

I use AI in marketing. A lot.

It can be brilliant for research, developing ideas, analysing information, organising a content strategy, finding gaps, improving structure and speeding up parts of the content process.

Pretending these tools don’t exist isn’t going to make a marketing strategy better.

But there’s a huge difference between using AI to help create good content and asking AI to create your entire voice.

If the process is simply “write me a 1,500-word SEO article about this subject” and then that article goes straight onto the website, you’ve skipped the most important part.

The business.

That’s where useful content comes from.

AI can help turn that knowledge into something structured and searchable. It can’t manufacture genuine experience you haven’t given it.

This matters for SEO too

For years, SEO content became associated with writing a certain number of words around a certain keyword.

Find the search term. Put it in the heading. Mention it enough times. Write 1,000 words. Publish another article next week.

Search has moved considerably beyond that.

People still use Google, but they’re also asking questions through AI platforms, looking at Google Business Profiles, searching social platforms and expecting useful answers without necessarily following the traditional journey from search result to website.

That makes genuinely useful, specific content more important, not less.

A business doesn’t need another generic article explaining “five benefits” of a service that has already been explained by 500 competitors.

It needs content that demonstrates knowledge of the work it actually does.

For a builder, that might mean explaining what happened during a difficult renovation and how the problem was solved.

For an interiors company, it could mean documenting a fit-out, explaining the decisions behind an installation or answering the questions customers repeatedly ask before requesting a quote.

For a financial business, it might mean taking a complicated question clients regularly ask and explaining it properly in language people can understand.

Those are useful pages for potential customers. They’re also much stronger foundations for search visibility because they contain something generic AI content often lacks: real context and genuine expertise.

More content isn’t necessarily better content

This is probably the biggest change AI has created for content marketing.

Producing content used to be relatively expensive. Someone had to research it, write it, edit it and publish it.

Now a business can generate 20 article drafts before lunch.

That sounds like progress, but it also means the internet is filling up with an extraordinary amount of average content.

So simply publishing more isn’t much of a competitive advantage anymore.

I’d rather build a smaller content and SEO strategy around genuine customer questions, strong service pages, detailed case studies, local knowledge and properly researched articles than fill a website with dozens of generic posts simply because we can produce them quickly.

The same applies to social media.

You don’t need another caption telling people your team is “passionate about delivering exceptional solutions”.

Tell me what you’ve actually done. Show me the project. Explain why you made a particular decision. Tell me something customers misunderstand. Have an opinion.

Give me a reason to remember which business wrote it.

So, should we stop using em dashes?

No.

Although I might ration them.

The em dash isn’t the problem. It’s simply become one of those things people notice because AI happens to use it rather enthusiastically.

And trying to fool an imaginary AI detector isn’t a content strategy.

The bigger question is whether there is a recognisable person, business or point of view behind the words.

AI is going to become an increasingly normal part of marketing. Used properly, it can make research faster, uncover opportunities and help businesses produce better content.

But the easier content becomes to produce, the less valuable generic content becomes.

Your experience is the bit your competitors can’t generate.

That’s what I’d build the content around.