I recently ran a quick LinkedIn post about human versus AI copywriting.
It was a bit of an ethical one around what percentage of artificial intelligence is acceptable in copy or content delivered from a professional copywriter (or content writer).
The sample was pretty [read: very] small, but here are the results:
- 26% said they agree with an emerging rule of thumb that seems to be 70% human and 30% AI.
- 32% said that they don’t care as long as the copy is good.
- And, coming in top at 42%, was 0% AI – 100% human only.
- No one ticked 50/50% split.
It’s interesting, well, it was for me anyway. And it got me thinking, what does each result actually mean? What’s the thinking behind the responses?
‘Human only’ copy – what does that mean?
Let’s start with the winner of the poll: 0% AI – human only copy.
Does this mean completely AI-free, including the process? Or that the copy needs to be 0% if you run it through an AI copywriting checker?
Is using AI okay for initial brainstorming?
For research?
For drafting?
What if a copywriter writes it and then asks AI to edit or rework it for flow – but doing that makes it seem 100% AI even though it isn’t?
Some of the staunchest supporters of ‘human-only’ copy are, in fact, copywriters. I get this. Writing isn’t just a job, it’s a craft we love, a process we’ve honed over the years. How dare a machine try to take it from us?!
Other sceptics are usually business owners or marketing managers who wonder why they should pay for copy if they can just use AI. “Surely I can just do it myself in minutes”, they cry.
But can they do the prompting, editing and shaping with the skill and nuance of a writing professional? Maybe a handful can.
Camp ‘as long as the copy is good’
Camp “I don’t care as long as the copy is good” makes logical sense.
Copy and content writing are services people invest in to achieve a specific outcome – whether that’s selling products or warming the target audience to the brand. If the copy works, what’s the issue if AI lent a hand, whatever the percentage?
The key here is that a generative AI copywriting tool like ChatGPT, Claude, or Co-Pilot still needs a human in the driver’s seat.
And that human – plus anyone signing off on the copy – must understand what ‘good’ copywriting really means. Otherwise, ‘good’ might just mean readable English, rather than copy that flows, builds a brand and drives sales.
The 70%/30% split
When I put forward the 70/30% split in the poll, I meant 70% written by the human, and no more than 30% written by AI.
I’ve seen some marketers flip this – 70% AI, 30% human – which is usually the approach high-volume agencies or content mills take when churning out lots of blogs or product pages quickly.
Neither approach is a hard-and-fast rule, but from a quality perspective, letting the human handle the bulk of the work usually produces more unique copy – a key indicator of quality.
Humans (read copywriters) bring judgment, creativity, emotional intelligence and nuance that AI still can’t fully replicate. We know how to choose the right tone, emphasise key messages, and structure a story so it resonates with the intended audience.
Relying on AI tools to create the full structure and draft is a bit of a cowboy strategy. Some of the outlines and drafts I’ve seen AI produce are a D- at best. Prove me wrong.
Artificial intelligence detectors aren’t all that
I briefly mentioned copy AI detectors before…the likes of QuillBot, Zero ChatGPT, and GPT Zero. These actually deserve a section of their own as they’re also a hotly debated topic.
The big question is, how accurate are they?
- Run a piece of copy through a few, and you often see a very different result in each. One might say 75% AI, the other 25%.
- Change a few words, and a whole paragraph can turn from AI machine learning to completely human.
- Use clear, concise English, and a well-known writing formula (that AI learned from writers originally!), and it can get called AI even if it’s 100% human written.
Because of this, AI detectors can’t be taken as gospel. They even admit it themselves in the small print. They can only try to figure it out based on formulas. So, never judge copy solely by the figures they spit out.
Copywriter vs generative AI tools ethics
Ultimately, AI itself isn’t the problem. It’s how copywriters use it. And how honest they are about it. That’s where things get murky.
If you’re passing off untouched AI output as carefully crafted, human-led copy, that’s shaky. Not because AI was involved, but because the client isn’t getting what they think they’re paying for. They think you’re a ‘writer’
But when they hire a copywriter, clients don’t just pay for words. They pay for:
- Perspective
- Judgement
- Clarity
- Decisions
- Creativity and original ideas
- Voice and tone
- Strategic thinking
- Emotional impact
AI can sound great, convincing, even, but that doesn’t mean it’s right. Just look at politicians! Strong human oversight knows what’s wrong and when something is a bad idea. This is what adds real value beyond words – it’s the skill you’re paying for.
If AI is being used, there’s a fine line between originality and efficiency that needs to be walked.
If you lean too far into AI, everything starts sounding the same (and we’re dulling our brains – that’s my greatest personal fear!). But when used well, it can actually sharpen ideas and expand creative thinking.
When it comes to disclosure, there’s no clear rulebook on this. But if you think the client would feel misled if they knew how something was created, there’s a problem.
The real results: human vs robot writing
There’s definitely increasing distrust around copy that is obviously AI. People are switching off. I know I do when I read another social post that’s come straight from ChatGPT.
That’s because AI-generated content often lacks the storytelling and empathy that drive real engagement. Without these, brand credibility suffers. Authenticity and nuance are what build trust. AI content can also miss originality and a distinct brand voice, leaving it feeling flat and generic.
The real data around this that exists at the moment shows a bit of a mixed picture.
Overall, it’s saying AI can beat or match humans for short, highly testable copy like ads. But human or human and AI-generated copy wins on higher-consideration, emotionally nuanced copy and overall ROI.
It’s not human vs AI tools; it’s both together
I think the answer isn’t human copywriter vs AI, it’s both working together. AI can speed up research, drafts, and edits, but humans are still the ones who shape meaning, emotion, and brand truth.
Saying 100% human content is ‘pure’ and the only acceptable option feels unrealistic. The real skill is balancing the two. Whatever the percentage, the ethical move is to be transparent with clients about how your copy is created.
Used thoughtfully, AI doesn’t replace creativity; it amplifies it. The real win is copy that’s accurate, persuasive and distinctly human, no matter what tools you used to get there. Will AI replace copywriters eventually? I hope not, unless it pays me a hefty redundancy package and I can retire early!
Want a copywriter to help you write copy that’s accurate, persuasive and distinctly human? Send me a message to arrange a free initial chat. I’ve been a copywriter since long before everyday AI existed.