ChatGPT, best partner in crime
I suspect I am not alone in being intrigued, confused and perhaps a little scared by the likes of ChatGPT and Bard. What does it mean for the work we do? For the jobs we have? Is it a threat? Is it an opportunity? I had read much about it, but wasn’t sure what it meant for me and my role as a creative and as a strategist, so I decided to conduct an experiment: I have spent the last few months using ChatGPT4 as a creative and strategic partner.
Admittedly, the first thing I did was ask 'Who is Saul Betmead de Chasteigner?' and found out I was no one, which was a tad disconcerting, that’ll teach me. But the more I used it, the more I came to understand what it could do, and how I could use it to be better at what I did.
The insights I gleaned are three fold: one obvious, one useful, and one surprising.
The obvious: getting the brief right
Its ability to get useful, curated lists of things from an infinitely wide range of sources was an immediate benefit - I did however often need to double check the accuracy. There were a few answers that just didn’t feel right, notably it re-wrote a few famous quotes which was odd. But it was the speed of response, the ability to test different thoughts, hunches and ideas very quickly that most impressed me, and I have never seen such a clear example of what most of us intuitively know: the tighter the brief, the better the answer.
This became most apparent when I gave ChatGPT a simple but challenging creative task on a subject with AI at its core: come up with new names for ships like those found in the late Iain M Banks Culture novels. The Culture is a very technologically advanced, post-scarcity, hedonistic society built around ships controlled by AIs called ‘Minds’. The names of the ships are one of the best things about these remarkable books, they kind of sum up the whole Culture and their irreverent attitude to most things. These names feel even more intriguing because these ships are also the most powerful entities in the universe, and sometimes vast, measuring over 200 kms in length and containing billions of people. They have names like Funny, it worked last time, Just read the instructions, and my favourites: Of course I still love you (a particularly powerful warship), I thought he was with you and Mistake Not…* and many more…
Firstly I asked ChatGPT to simply come up with some new names, and it created things like Galactic Explorer, Infinite Horizon, Quantum Odyssey, and Cosmic Harmony. Ouch.
I pushed it for names with more humour, and it created The punny platypus, The goofy gorilla and The jolly jellyfish. I don't need to go on. Things got much better when I got much more specific. I asked for names like Funny, it worked last time and it gave me Sorry, I am not a morning ship, Winging it, Slippery when wet. Improving, but there were still some ones you'd rather forget: Awkward turtle.
I then told it that I preferred the first two and asked for more like those: Making it up as I go along, Not my department, and It wasn't my turn to watch him. It was getting closer to what Mr. M Banks might have considered and to what I, an avid reader, might like.
The useful: knowing where not to play
The Culture Ship name exploration also highlights another insight: you don't have to play around for very long with ChatGPT to realise it is incredibly good at identifying the mundane, the commonplace, and the obvious.
Its ability to create an exhaustive list of the “things you shouldn't try” is fantastic. It’s because of this, you then quickly know what not to do. Just ask it “to give you creative ideas to launch a new electric vehicle” or “to give you radical innovations that will revolutionise how fine art is owned”, and you'll see what I mean. When I tried these sorts of requests, it either came up with things I might consider but then reject immediately, or were so vague they were only helpful to capture vast territories I could then explore properly.
The last point is worth unpacking further: ChatGPT quickly provides you with a list of very broad territories, it also gives you a list of things to try and recombine to find something genuinely distinct and useful. Given recombining what already exists is a pretty good definition of creativity, it’s a particularly effective weapon if you are in the business of innovative ideas.
The surprising: an effective work partner
So we have a technology that works best with a clear brief, and because its feedback is almost immediate, it is a valuable way of quickly fine-tuning, and getting to a better brief. Also, it is very good at finding exhaustive, foundational territories, which you can either discard, build on, or combine in novel ways to find something special. Combined, you get something very handy indeed.
ChatGPT has ended up feeling more like a partner in work than anything else. I could ask stupid questions I thought I should know the answer to. I could bounce ideas off and see where they went without consequence, without judgement. And I could do this fast. Strangely, I felt I had more permission to push my ideas and felt more comfortable in the making.
For those of us who sometimes work on projects alone, or often have to go away and think things through in isolation before coming back and sharing, this can be remarkably helpful. I have found myself using it more and more to refine and then explore, from organising convictions for businesses, to brand and communications ideas, NPD, marketing approaches and everything in between.
Does it make those really impressive leaps, the ones that just feel right and where it isn’t quite clear where they came from or where they might go? Nope. Can it lay the foundations for those leaps? Definitely.
*Full name of this particularly powerful warship: Mistake Not My Current State Of Joshing Gentle Peevishness For The Awesome And Terrible Majesty Of The Towering Seas Of Ire That Are Themselves The Milquetoast Shallows Fringing My Vast Oceans Of Wrath.
Photo by Reza Jahangir / Unsplash