Will Copywriters Be Replaced by AI?
Predictions for the future of copywriting
Are you wondering if copywriters will be replaced by AI and intelligent machines? Maybe you’ve never given it a thought. Maybe you’re counting down the days.
Will copywriters be relevant in five years time? Will automation make the economy tank so hard nobody buys anything anymore?
I’m no oracle. The art of futurism is often judged simply by how persuasive you are. People ignore bad calls while celebrating wins — confirmation bias at its finest. But I reckon I know a few things about machine learning, AI, copywriting and the changing nature of the global economy, so let’s have a look.
1) The global economy will hit a recession, but creative capitalism will not end.
Things never stop moving. Almost any economist will tell you a global recession is coming. But it’s not the end of the world. Things will always need selling.
The world is a swirling whirlpool of clashing, changing forces. Despite the saturation of fear in the news (the most profitable of emotions), there’s an overwhelming abundance of creative good happening in the world.
Engineers are making clever gizmos to take plastic out of the oceans. Hungry entrepreneurs are designing innovative products despite economic anxiety. Brave survivors of abuse are addressing injustice and the balance of power. These give me hope.
Solved problems create different problems. The mobile phone, opening up lightning-speed global communication, caused so much lost sleep that further innovation had to appear — so now the colour temperature on phone screens changes through the day. Alongside this, a whole new anxiety economy sprung up, so we can pay £9.99 per month for a meditation app on our phones to help us de-stress from looking at our phones too much.
2) We will always consume information by the written word
There are only so many ways we can take in information — at the moment. Images, printed words, digital words, speech, video, inter-personal, and so on.
Have you ever searched for knowledge on Youtube and found yourself cursing a video’s 10-minute length for what seems like a simple query? You madly skip through the timeline to get to the point while the creator is denied monetisation because viewers aren’t looking at their ads enough.
It’s the same reason we set podcasts to play at 1.25x speed. There’s often an asymmetry between how information is delivered and how we want to consume it. This is why, if I want to know something, I’ll usually look for an article rather than a video.
And search engines can only really search text. Until they understand semiotics, tonal inference and abstract concepts, they won’t be very useful in searching non-textual media. Irritatingly, I have to edit my subheadings in this article so a dumb machine can understand them rather than show respect to my readers who can understand the journey I’m taking them through. Argh.
This is why we read books. Being exposed to ideas in long-form means our consciousness has time to wander around them, exploring them from different angles to fully understand their nature.
People won’t stop reading, no matter how fancy videos become.
3) Methods of information delivery will change
Technology is driving us towards an expanded array of information media. Let’s look at education.
I’m a visual learner; I have trouble remembering some abstract concepts. Finance is difficult — not so much mathematics, but economics. I can only remember the difference between Net Profit and Gross Profit by imagining them as different locations on a balance sheet, written in different colour inks.
I can understand scientific concepts like physics — where even multi-dimensional quantum weirdness can manifest as waves and particles bobbing around in a galactic soup. But where do I even start with Credit Default Swaps, Modern Monetary Theory, and the Inflationary Debt Cycle? What do they look like?
This animated video, narrated by affable billionaire Ray Dalio, made that last one a bit easier (unfortunately it presents an economic theory as the economic theory, but it’s still worth a watch).
But imagine if I could use a virtual reality application that showed me 3D blocks floating in front of my eyes, representing monetary instruments, that I could interact with in real-time. I could watch income turn into debt turn into surplus, and actually understand the flows of something that exists not in the material world, but only in our collective imagination.
There are legions of people around the world with different learning styles and abilities to understand things. The tech is already there. Just needs someone to build and deliver it.
That’s education — how about marketing?
4) Copywriting will be augmented by technology, not replaced — yet
Persuasion by non-textual methods will grow, but persuasion by the written word (copywriting) will always be a rapid, efficient method of steering behaviour that can be indexed by machines and subtextually understood by humans.
Machine learning is now being used to contribute to copywriting, by enterprising engineers who understand that language is made up of logical blocks. This is why I think of copywriting as a science more than an art (although the art is still very much there). It’s a highly structured form of linguistics, informed by years of psychological results-driven research. Anything this highly structured is always going to be vulnerable to automation of some sort.
Here’s an example: Markov Chains — random processes ‘trained’ from ingesting real data. The kind of thing that prompts autocorrect on your phone. When you type “copy”, it might offer up “copying” or “copyright” as the next suggestion. If you regularly send emails about copywriting, it’ll likely offer “copywriting” as the next option. That doesn’t mean it understands the meaning of the word — just that you, the writer, normally type it in such a circumstance. It does not understand what copywriting actually is.
Likewise, search engines, as clever as they are now, don’t really understand the text they index. If I was a competent marketer, I’d want this article to rank for “AI Copywriting”, or “will copywriters be replaced by AI?”
That’s probably the best way people would find this article. But it looks absurd to sprinkle those phrases through the text. We don’t need to explicitly label everything we write like that, because as humans, we understand context and subtext.
5) Machines don’t understand the nuances of copywriting
It’s simple to run a structural analysis of a sales page. This is why you can find sales page “formulas” in copywriting guides. Something like, Attention > Offer > Features > Objection Handling > Social proof > Testimonials > Urgency. They all lead the reader on a journey towards the sale.
Think about positioning a price point. We’ll use an example of an industry where most differentiation is financial — insurance.
Rather than saying,
“You’d save £200 per year by switching to us”
a savvy copywriter would swap it around:
“You’re losing out on £200 per year by sticking with your current provider.”
This utilises the theory of loss aversion (where a loss is seen as more painful than a gain of the same amount is pleasurable). Would a software system ever understand this theory enough to apply it correctly?
It’s a simple structural swap in our minds — but there are too many variables, nuances of language and psychology, to really have this automated.
This makes it really interesting to consider whether AI could ever come up with a brand. I don’t mean a logo and tagline — but an actual brand with values and a story, able to attract people to its cause in a way they might not be able to explain. After all, a brand exists only in the imagination — it’s utterly intangible.
My gut feeling points to no, it couldn’t. But technology isn’t going to stop. Innovation causes the unimaginable to become real.