There’s been an enormous upheaval in the digital marketing industry over the last 5 years or so, and in the broader economy, too, as companies attempt to roll out AI in order to save costs on human capital.
This, along with other environmental and ethical concerns (to name just two more, there are many) has made AI, specifically generative AI, something of a highly contentious issue.
My field has also been affected. Generative AI is capable of producing blogs, briefs, drafts, summaries, and even prose and poetry with speed and facility that human writers simply cannot match.
I say this honestly, and humbly. I cannot match the speed of a generative AI agent. But there are things that I, like other human writers, can do, and which AI agents and generative models cannot.
These are some of the biggest among them, some of which still have impacts on eCommerce copywriting, engagement, conversions, and more, all of which affect SEO, PPC, social media, and other digital marketing campaigns.
Human Copy Can Be Truly Original
AI needs human sources to generate copy in the first place. Whether you’re asking a generative AI model like ChatGPT to write a sonnet or draft a blog on how to start a fire with a ferrocerium rod, it simply cannot do it on its own.
It needs other sources to scrape, and an AI model is only as good as its data set.
Now, in a sense, a human writer is the same. You can’t ask a baby how to tie a shoe. Humans, like AI agents, must be taught. But it is also true that human experience must necessarily predate AI copy because humans created AI.
Not to wax philosophical here, it’s just that my point is that everything that’s ever been generated by an AI agent before was first an idea in the head of a human, somewhere.
The ultimate conclusion of all this is that leaning too heavily on AI will eventually make the internet one big echo chamber if we’re not careful.
That’s the number one reason that human-written copy is still effective. It can still be truly novel whereas anything produced by AI is necessarily a recreation rather than a creation.
Human Copy Can Be Experiential
In the same vein that human copywriting can be entirely and truly novel, it can also be experiential, and that’s something that AI cannot do now and necessarily be able to do.
AI can’t talk about strategizing for long-term growth for an SEO campaign. AI can’t write a true review of a pair of boots. AI can’t come up with a good way to replace store-bought leather conditioner with a homemade alternative through trial and error. It cannot do these things because it has never run an SEO campaign; it has never worn a pair of boots; it has never conditioned leather.
Now, as I stated, it can take what others have already done and repurpose them, but the synthesis of a summary is not in and of itself a true experience. Not a novel one, at least.
Human Copy Has Quirks and Emotional Depth
I hate to quote Justice Stewart’s famous phrase, because it is nonspecific and therefore makes for poor definition, but, as for human copy, “I know it when I see it.”
What I mean by this is that, sometimes when I’m reading something, it’s just so unique in its voice, the idioms the writer uses, or the expressions and anecdotes the writer is conveying that it’s almost impossible that a generative AI model cooked it up.
As an example, check out this article recently published by the Marginalian on GK Chesterton’s ponderance of the dandelion, the fundamental wonder of existence, and the blessing that gratitude is when we give it to ourselves as a gift. I dare not profane Chesterton’s words or the brilliant prose with which this particular contributor has graced The Marginalian in this piece. If you are interested in it, read it for yourself.
But, just the intro: even the expression “rosary of days”; it’s just not the sort of thing that generative AI could or even would come up with if it were capable. There’s a beauty in the mundane that inspires human writers to select their words, and which AI cannot grasp, that isn’t entirely about communicating an idea but an emotion.
Human Copy Can Be More Poetic Without Descending into Clichés
AI does a great job of speaking in platitudes and it seems to understand poetic meter, but it doesn’t write anything really artistic that doesn’t seem templated. That means, for all intents and purposes, what it creates is art in the strictest sense of the word because it has been created, technically, but beyond that, AI doesn’t seem to have any soul.
I’m not going to beat this dead horse, because it’s not something that only I have noticed, and it’s also something that’s fairly hard to describe until you spend a lot of time reading AI-generated text. Once you do, you’ll understand what I mean. Too many em dashes, cliches, blunt metaphors and sentence fragments typify AI-generated art and they don’t make for good substance.
Humans Are Better at Drafting Conversion-Oriented Copy
I’m actually not sure what the reason is for this, but there have been studies that suggest that where conversions are the goal, human copy outperforms AI-generated copy, and by no small margin.
Therefore, if your goal is to get people to actually read what your copy says and convince them to complete a purchase, sign up for something, download a form, or anything else of that nature, your bets are best with human writers.
What AI Does Better Than Human Copywriters
Now, granted, there are some things that AI can do better than human copywriters, so realistically, I can’t sit here at my computer and say that eCommerce businesses should shun AI entirely because they really shouldn’t. Someone like me has little use for it but that doesn’t mean there’s no value there at all.
So, in the spirit of fairness, consider the following:
No matter how you look at it, generative AI models are faster than human copywriters, even if what they produce is objectively low in quality. Sometimes quantity is a quality all its own, and human writers, fast as they might be, cannot keep up with AI.
AI can summarize vast sets of data much more quickly than human copywriters.
AI rarely if ever makes grammatical and clerical errors. Typos are effectively if not entirely nonexistent. So for proofreading purposes, AI does much better than a pair of human eyes.
Anything that has to do with production at scale; a single AI agent can do the work of entire teams of people in a small fraction of the time. This goes beyond copywriting, by the way.
If you’re getting the impression that AI mostly takes the cake when it comes to large-scale synthesis, summary and generation, that is the case. The issue really comes down to soft signals associated with brand voice, artistic originality, and other market cues, like what readers or shoppers prefer. Humans still do better with that.
So, Which Makes Sense?
At the end of the day, there is a place for both human writers and AI agents for most businesses. It’s just a matter of being able to accurately identify each of their strengths.
We use both here, for different things. For your purposes, AI models may be able to help you parse data, come up with drafts for pages, and do coding and design work. When it comes to looking everything over or coming up with the ideas in the first place, well, that’s where you still need humans.
And, of course, if you’re restricting the scope of where you should use AI to copywriting, anytime the subject matter is nuanced, esoteric, or requires firsthand experience, go with a human writer that knows the subject every time. You can use the AI agent for proofreading, but don’t go past that.