ChatGPT - AI has landed
The future is here… I think we can all feel it. And those of us in the creative industries can feel an uncomfortable encroaching sense of being shuffled on.
I definitely felt this way when I first heard of ChatGPT and how it had launched into the world, ready and set to take over the world of writing. Initially, I was dismissive, then a little dismayed and finally a little scared. Did this mean that the craft and skill of writers would be put aside for this bot that could write a 1000 word blog in less than 2 minutes (and for free)?? Did it mean that, essentially, the thing I loved to do and was good at (and being paid for) would be defunct? Not needed?
I didn’t go near the thing for fear that I would be proven right. But…
Then I started to talk to people. I talked to my corporate husband (who obviously wants me to keep earning), and he used it for emails or to smarten up some of his communications. I talked to other writers, and they use it as a tool to convert language from casual to corporate. I spoke with designers, and they’ve already felt the encroaching future with the advent of AI design or even the likes of Canva. I read articles about the tool and listened to the ‘experts’ (amazing how quickly people claim themselves experts, after a few days of experimenting). And finally, I listened to a podcast by Kate Toon on ChatGPT and got over myself.
I jumped onto the free version and started ‘playing’. Playing with prompts. My first attempts were fairly dull. I laughed at how lacking in personality and formulaic it all felt. I suggest trying to write a poem, I asked for a poem about a dog, unicorn and truck. It was very sweet, but each line rhymed, and there were four lines per verse… surely we humans can inject more?
I’ve since done more research, and it does come down to how your prompt and ‘speak’ with ChatGPT. Briefing the tool the right way can come out with some very different results. Just like with a human, the more detail, constraints and requirements you can give, the better the output. But all the ways I tried, the results were still a little bland, and I was vividly aware that all of the words had been pulled from websites, often verbatim. Don’t get me wrong; often when I’m writing for clients, I do a wee tour of what ‘is already out there’ - read as much as I can, form an opinion and write based on the tone of voice, my knowledge of my client and their business and what I have read from other sources.
ChatGPT and all those AI tools that follow will threaten the field of writing. No doubt about it. But I don’t yet feel there is no place for a human to write for humans. ChatGPT will develop, learn and grow the more we use it and will become wonderful at doing what it does. But again, with all of these tools, the output is only as good as the input.
There is a place for writers in this world, our nuances, thinking and ability to pull together thoughts with flourish, thought and humour. And frankly, the clients I work with realise that the human element is what is important. Maintaining that we are writing for humans, not for Google (only) or for other AI bots. Bringing together the information, nuances, and experience will win on the day—and genuine human warmth.
There is a place for these tools, don’t get me wrong. We are running a race to keep up with the technological advances of our time. And utilising these tools makes it all the easier. I use Grammarly, it’s set on my laptop and constantly makes suggestions on all of my written comms. Do I always agree with Grammarly? No, not at all. Sometimes the suggestions misconstrue my point or are far too formal. But do I like the safety net? For sure. And some suggestions are appreciated.
ChatGPT and now Grammarly has their own version coming out - are here, and they won’t be going anywhere. I’ll learn how to use these tools. See how I can weave them into what I do. But it will always be with an eye on how they serve me and my clients. I will always write for humans and intentionally for what is interesting to read. So nothing really has changed. But I might use ChatGPT to get past my blank page when I’m stuck or even to look at a different way of wording something. But (for now), I genuinely believe that humans can tell if it’s been auto-generated. And humans are what matter.