Human vs Machine: Can AI really replace human copywriting?

WRITTEN BY

Rebecca Nordqvist

PUBLISHED

15th November, 2025

TIME TO READ

6 minutes

Please read it. But here’s the gist.

AI is rapidly taking over how ads are delivered, but it still struggles with why people care.

While platforms like Google Ads can auto-generate headlines and optimise for conversions, they often miss the emotional nuance that makes copy truly connect. Our research shows that AI-written ads can perform reasonably well, but human-written copy still wins when it comes to empathy, values and aspiration.

The takeaway? AI works best as a support tool, not a replacement. The strongest results come when humans set the strategy and emotional intent, and AI helps scale and optimise the execution.

If you’ve logged into Google Ads recently and thought, “Hang on… who wrote that headline?” – you’re not imagining things.
AI is quietly (and sometimes not so quietly) moving into the driver’s seat of digital marketing. From automatically generated ad headlines to ‘helpful’ recommendations designed to boost conversions, platforms like Google Ads are increasingly saying:

“Don’t worry, we’ve got this.”

- Google

But anyone who’s actually run campaigns knows… it doesn’t always work the way it promises.
So that begs a bigger question:

If AI can optimise bids, audiences and copy – do we still need human copywriters at all?

This exact question formed the basis of a research project I completed as part of my Masters of Digital Communication. And instead of debating it theoretically, I decided to put AI and humans head-to-head where it really counts: Advertising headlines that are meant to make people feel something.

Let’s unpack what I found – minus the academic jargon, and with a bit more real‑world context.

The chef and the oven 🍳

Think of digital marketing like a professional kitchen.

AI is the industrial-grade appliance:

Human marketers are the chefs:

Now imagine the oven deciding it knows better than the chef and changing the recipe mid-service.

That’s essentially what we’re seeing with platforms like Google Ads auto-generating headlines, combining assets, and optimising for conversions based purely on data.

Sometimes? It works beautifully.

Other times? You end up with ads that technically follow best practice… but feel cold, generic or just off.

Why human connection still matters (even in automated ads)

Yes, online advertising is becoming largely automated now.

Bidding, targeting, delivery, optimisation – it can all be handled by machines.

But copywriting sits in a slightly awkward space.

Because great copy isn’t just about clarity or keywords. It’s about:

These are messy, human things.

AI is excellent at spotting patterns. Humans are better at understanding meaning.

And when you’re selling something tied to identity, self‑esteem, health or values (hello cosmetics, wellness, education, services…), that difference matters.

How we measured “connection” (without guesswork)

To keep this grounded, I used the AIDA framework, a classic marketing model that maps how people respond to advertising:

For this study, the spotlight was on the last two stages — desire and action — because that’s where emotional connection (or lack of it) really shows up.

The experiment: AI copy vs human copy

Participants were shown two Google-style ads for the same cosmetic product:

One written by an AI copywriting system

One written by a professional human copywriter

No labels. No clues. Just the copy.

They were asked:

The results: not as black and white as you’d think

At first glance, the outcome looked decisive:

👉 70% of participants preferred the human-written ad.

Case closed, right?

Not quite.

When participants rated desire and action on a scale, the gap between AI and human copy was surprisingly small, roughly 5–12%.

Translation:

This mirrors what many marketers see in Google Ads today. Auto-generated assets don’t usually tank performance, but they rarely elevate it either.

So what was the real difference?

Where AI and Humans Actually Diverge

Once participants explained their reasoning, four clear themes emerged.

1. Readability: clear vs considered

AI copy tended to:

Human copy tended to:

Some people loved the directness of the AI. Others found it robotic or rushed.

This is exactly what we see when Google auto-combines headlines — technically fine, emotionally hit-or-miss.

2. Ethical language & the buzzword problem

Words like ethical, natural and organic showed up in both ads.

But perception varied:

AI used these terms correctly — but safely. Human copy provided context and nuance.

That subtle difference had a big impact on trust.

3. Aspiration vs assurance

AI copy focused on outcomes:

Human copy leaned into aspiration:

Neither approach was wrong, but aspiration triggered stronger emotional responses (both positive and negative).

AI doesn’t yet know when a phrase inspires… or when it accidentally alienates.

4. The ‘but I still gave questions’ factor

Across the board, participants wanted more concrete information:

Pricing. Ingredients. Proof.

Which highlights an important truth:

Emotional copy works best when it’s backed by rational clarity.

No amount of optimisation can replace that balance.

So… can AI create human connection?

Short answer?

Yes — but not instinctively.

AI can:

But it struggles to:

Which is why auto-generated ads often feel acceptable — but rarely memorable.

What this means for marketers (and why this isn’t bad news)

This research doesn’t suggest humans are being replaced.

It suggests roles are shifting.

The most effective setups going forward will look something like this:

In other words, AI becomes the appliance, not the chef.

The takeaway

AI copywriting tools (and platform-led automation like Google Ads) are powerful.

But connection isn’t created by optimisation alone.

It comes from understanding:

For now, that still requires a human in the loop.

And honestly?

That’s not a weakness of AI, it’s a reminder of what good marketing has always been about.