By Vikki Maver

AI-generated copy and content is getting convincing. Too convincing. Emotive words and phrases? Tick. Balanced and methodical sentence structures? Tick. Catchy headlines and CTAs? Tick… and tick.

So I don’t blame the budget-strapped marketer for thinking: Why the hell are we paying for copywriters? Why carry that cost when AI can churn slick words out in seconds?

And the numbers bear that out. A recent Wall Street Journal report found that 36 percent of CMOs expect to cut jobs due to AI in the next 12 to 24 months. Meanwhile, the demand for freelance writers had already dropped by over 30 percent by the start of 2024. 

The pressure for marketers to bake AI into budgets, workflows and hiring decisions continues to build. And when it comes to copywriting, the justification is easy: if AI can produce copy, then the people who produce that copy are redundant. The danger here is that marketers are working off the (very wrong) assumption that the role of ‘copywriter’ is just about execution, overlooking the strategic thinking and expertise that underpins it.

As the founder of a copywriting agency, I have an obvious stake in this argument (and you should factor that in). But here’s what I can tell you from the ground: we felt the shift early. I still remember losing our first long-term client to AI in late 2023. I was disappointed, yes. But I was also curious: would they be able to tell the difference between our work and the AI-generated mimicry they’ve replaced it with?

Three years on, they haven’t called to say so. But what I do know is this: the drop in quality is glaring to me and my copywriting team. And if the data is anything to go by, the people who matter most – their audience – can see it too.

While 82 percent of in-house content teams believe AI-generated work matches the quality of human writing, 52 percent of consumers reduce engagement when they believe content is AI-generated. And brand trust drops accordingly. 

This means the gap between what marketers think is working and what audiences respond to is widening. But because brand trust is rarely measured alongside day-to-day performance metrics, that gap largely goes unnoticed – and therefore unchallenged.

This is the Dunning-Kruger effect, marketing edition. The Dunning-Kruger effect describes a gap between confidence and competence, where the least skilled are often the most certain. And that’s exactly what’s going on here. 

Teams trust AI-generated copy because it’s polished and technically correct. It looks ‘right’. What it lacks is harder to name and easier to overlook. That’s why the feedback loop continues: content is generated quickly, approved without friction, published and performs at an acceptable level.

The problem is, ‘acceptable’ is a low bar. And a dangerous one. Because by the time the impact shows up – in a loss of brand trust and distinctiveness – it’s hard to trace it back to the content itself. So the machine keeps moving. Just with diminishing returns.

The tools aren’t the issue here. Used well, they raise the baseline and make content easier to produce. The risk is treating them as enough – letting them replace the need for judgement altogether. And that’s why the skills of a copywriter now matter more, not less.

Experienced copywriters know how to interrogate a brief, make deliberate choices about what to emphasise and what to leave out. They also know how to find the angle worth pursuing, the tension that makes someone lean in and the phrasing that moves them to act. And, above all, they know when AI output misses the mark.

AI has raised the floor on content quality. That’s genuinely useful. But it’s also created a generation of marketers who can’t distinguish between copy that’s polished and accurate – and copy that’s strategically sound and credible.

The brands that win will be the ones that can still tell the difference… and invest accordingly.

What does this look like in practice?

If AI is part of your workflow, it needs to be supported by clear standards and stronger judgement. That means being more deliberate about how content is created, challenged and approved. Here are a few ways to put that into practice.

1. Get clear on the message first – always: The best copy starts with a sound strategy – and a clear idea on what your brand should be saying in the first place.

2. Pressure-test the idea: If the core message could apply just as easily to a competitor, it needs to go.

3. Watch for AI clichés and generalities: Strip out overdone phrases and sentence structures, and words that say very little.

4. Slow down approvals and interrogate the output: Never accept output at first glance. Push it until it proves it works. Then decide if it’s worth publishing.

5. Build copywriting capability in your team: Most content people writing with AI have never been trained in the fundamentals of good copywriting. It’s never too late to fix that.

6. Keep expert oversight in the loop: Make sure the person approving your content has the right experience and expertise to judge whether it works.

Vikki Maver is the founder and lead trainer at Communication Skills Academy

Read more: How to create content AI can’t replicate