Will AI replace copywriters?
Will studious, hard-working AI like this replace me?
AI writing tools have been the talk of the town among copywriters and content marketers lately.
Before, this technology was almost universally derided by writers. And this was completely justified, as the AI produced writing that was laughably bad.
However, there’s no denying that AI writing tools have advanced significantly recently. GPT-3, an AI model created by OpenAI, has been trained on vast amounts of text, amounting to a staggering 45 terabytes of data. It then uses this information to produce content that’s written in a way that mimics human language.
In recent days, the media has been abuzz with commentary about how ChatGPT - an AI chatbot, also created by OpenAI - is capable of moments of originality and brilliance. When AI can pen poems in the style of Shakespeare, you know something pretty remarkable is underway.
As a freelance B2B tech copywriter, I’m always excited about the potential of innovation - that’s why I specialise in this space. But there’s one key reason to be sceptical that AI will replace professional copywriters that, amid all of the frenzied discussion, hasn’t been commented on nearly enough.
AI writing tools don’t properly fact-check
If you’re looking to produce content that provides high-quality information to your audience, AI tools fall at the first hurdle due to their inability to verify what they’re writing.
This is because, as mentioned, the tech works by reading everything that’s out there on the internet and basing the content on that. So this means that if there’s a lot of unreliable information out there on a subject, that’s going to be the basis of the writing. And as everyone knows, the internet is full of just such information, or worse still, actual misinformation that intends to deceive or contains harmful biases.
And even if most of the content out there on a particular subject is accurate, it only takes one or two small inaccuracies to produce an embarrassing result. For example, a computer scientist who searched for her own name in ChatGPT received a mostly correct answer, apart from one small detail - she was informed she’d passed away. OpenAI themselves conceded that the tool can produce “plausible-sounding but incorrect or nonsensical answers”.
If you ask the AI to create content on a really innocuous subject that doesn’t require any rigour to cover, and where objective truth isn’t really at stake - something like ‘Why labradors make great pets” - you’re probably not going to encounter any of these issues.
For companies who want to produce content like this, AI could be a useful time-saving tool to have in your arsenal. But at least in my niche - B2B tech - the ability to guarantee the information you’re providing is 100% accurate is critical.
In B2B tech content marketing, we write about complex subjects where claims need to be backed by high-quality data and research. If you fail to do this, you’ll lack the kind of credibility that’s necessary for your audience to trust you and for you to become an industry thought leader.
You can’t replace human critical thinking
Ultimately, we’re a long way from creating AI that has human-like powers of critical thinking. In the context of content marketing, this means that we won’t be able to rely on AI to have the ability to smell when something just isn’t right with a particular source.
So tools like ChatGPT could certainly help writers like me in certain contexts, for example by summarising and condensing a block of text I’ve already written into a 100-character meta description. In a sense, this makes it similar to a tool like Grammarly: useful in some situations, but not something you can rely on exclusively.
It’s going to be interesting to see how this technology continues to evolve, but I think writers who have the intellect and critical thinking skills to produce high-quality, rigorous copy will be in demand more than ever in the B2B tech space.