We all depend on copywriters to create compelling headlines, adverts, campaigns and other content that gets the attention of our audience. But what if AI could do this instead? It’s a valid question because artificial intelligence is getting better every day at producing human-like text and delivering marketable messages.
It’s also a question that has been asked (nervously) many times before! Richard Feynman, the great physicist and Nobel laureate, said there was “a tremendous future for AI as long as it can be used to help people”.
I am sure you have all watched sci-fi films like The Terminator, and wondered when ‘robots’ will take over our world. Will it start with copywriting, let’s find out 🙂
Why Are AI Writers Such a Threat?
At a time when marketers are struggling to find enough writers, particularly in digital and social media, it is tempting to think that smart AI can be brought it to save the day.
If AI really could take over the jobs of a human copywriter, the impact would be huge and change the industry forever.
Why? Because AI can generate posts and articles much faster than humans can. When the AI gets smart enough, It can construct long-form copy with words it hasn’t previously seen or learnt about, which you would struggle to do by yourself at such a rate.
But Is the AI Technology Ready for Mass Copywriting??
At first glance, it seems like we are almost there with AI. We already have AI that can write simple music and short stories, and now researchers from Stanford University have a system called ‘Poetica1’ that produces poems.
So why can’t these same systems be applied to copywriting?
Well, they probably can… but there are some major challenges that we will talk about later in this article.
But first, how do the popular AI models and algorithms work?
AI (short for ‘artificial intelligence’) is a type of machine learning that makes predictions or decisions.
One type uses ‘supervised learning’, which uses algorithms to analyze data sets and learn patterns from them. After a computer has been trained on thousands of data sets, it will react and provide a solution similar to what the real answer might be.
To make this simple, let’s say that you teach an AI system to play chess.
The computer is given some chess boards and data around the successful moves made in past games. It is then better able to predict the next move or get out of sticky situations based on what it saw before. In this case, the AI will be “playing” against different players (i.e. solving different problems) in order to improve its skills. However, you have helped it understand the game that it is playing in the first place and tagged data appropriately.
Another kind of AI is one that uses ‘unsupervised learning’. This is where you feed the computer with lots of raw data (i.e. text or video) but don’t tell it what to do with it. In this way, the computer has to find its own solution based only on the raw material. It can essentially find patterns and trends in data without any outside help.
Most commonly used AI today are in the supervised category, but we expect to see a gradual shift to more unsupervised AI as the technology develops.
This is relevant to copywriting because we are currently only using supervised learning. In other words, the computer is taught what to do by humans (the tagger).
What are the Key Problems with AI Copyrighting Right Now?
As we are mostly dealing with supervised AI, we are dealing with something far less intelligent than a human copywriter
It is only capable of finding trends and patterns in the data set it is fed. These AI systems are not capable of understanding the meaning behind words, so they can’t create content that will consistently connect with your audience and spark an emotional reaction.
In short, these AI are not yet capable of creating unique and compelling content.
For example: if you feed a computer with enough data from past successful ads, you can teach it to create duplicate adverts. These will probably sell as well as the real thing, but they won’t provide any additional value to your product or service because the AI has just been fed the same information again. It hasn’t learnt anything new or different to what it already knows work in “real life” just yet.
Put simply, an AI doesn’t know the intricate details of the world we live in, so they can’t use this knowledge to add to and improve the copy they create organically.
Innovation and creativity are a massive part of copywriting, and the current set of AI tools doesn’t have algorithms smart enough to provide this yet.
So no, they cannot take over from human copywriters just yet. However, these AI writing tools do provide excellent assistance for human copywriters, helping them to brainstorm ideas quickly (for example). For me, this is the main use case for AI in copywriting right now.
To be honest, I am amazed at how AI in writing has come on over the years. First, we had AI based grammatical and syntax checkers that told us if what we wrote was accurate or not. Now we are getting to the level where AI can assist us and even start adding its on sections to the content we create. Can it make the next step to fully created AI content? We are not there yet, but who knows in the future with the rate that technology is moving on.
To Sum Up, This is Why Humans are still Needed in Copywriting!
The human touch is needed to provide full context to what is being created, something that an AI cannot do. Only people know about the intricacies, the personalities, and the little details that will make a piece of content worth reading. Even if AI could create fully realized pieces of content from scratch (it can’t yet), it wouldn’t be able to use them in context. It would have to learn these things for itself, killing off any efficiency gained.
They’re still a lot of errors and inconsistencies in the copy that AI creates, so a human is always needed to vet what has been created and polish it up. If you let a fully AI copywriter on the world unchecked, it would be chaos in no time. AI won’t know what is needed, what is important, and what makes good content (at least yet it cannot) so this alone would make it a disaster for marketing.
The recent developments in AI technology with GTP-3 have seen it taking another step towards replacing human copywriters, true. But the next step is a truly massive one and will take time to complete (if at all). So, if you’re a copywriter, no need to throw in the towel yet. Plenty of good years in you yet 🙂 Just be ready for a fight somewhere down the line!
Plot twist: I used an AI writing assistant called ShortlyAI to help me write this article 🙂 The Robots are coming….. ha ha 🙂