I must admit, when mainstream generative AI reared its ugly head (and deformed hands) approximately two years ago, I was deeply concerned for my profession and my job.
It began slowly, with a pithy haiku here, or an image of a Pokémon ‘in the style of Van Gogh’ there, until it snowballed into high fantasy art, short stories, and photo-realistic imagery that has become difficult to differentiate from the real thing. People outside the design space were quick to point out this new technology’s encroachment on ‘our turf’, and were even quicker to ask, in an almost gloating way: “Aren’t you worried?” and “What are you going to do now?”
The answer, invariably, was always “Yes, I am worried,” and “No, I have no idea what I’m going to do.” Then the media train whirred into gear and we were all being told to “skill up or step aside” or “The race to master these tools finished yesterday,” which only added to my “designer angst” and Weltschmerz. That was until I found the time to actually sit down, work with these tools, and think critically about this new technology and what it means for me as a designer, and for us as a society.
Through interacting with these services, I have slowly come to the realisation that generative AI is really only exceptional at answering ‘definitive’ questions. By ‘definitive’, I mean questions that have one, agreed-upon, established answer. 1+1=? is a definitive question (philosophical constructivism aside) where most people on the planet would agree that the answer is always going to be 2. Moving your pawn to E4 on the chessboard is a definitive question, albeit with a wider but finite set of definitive answers (that is, if you want to win). “Reimagine me as a Sesame Street character” and “What is our current quarterly revenue?” are two more definitive questions. All of these definitive questions have a concrete answer that can be verified. Generative AI excels at ‘definitive’ questions, and as such, has already begun an upheaval (similar to what we designers experienced two years ago) in the ‘definitive’ question realms of development, accountancy, medicine, and even law.
The shoe is now on the other foot.
Why? Because ‘indeterminate questions’, the antithesis of ‘definitive’ questions, are precisely the domain of traditional design and creativity, and an area where generative AI performs exceptionally poorly. If you've ever tried to generate a truly specific and original idea (image or text-based) using generative AI alone (note: “original” is not a photo of you and your partner as a Studio Ghibli pastiche), you likely experienced a frustrating cycle of wasted tokens and prompts, often ending with a less-than-satisfactory result. You might settle for “good enough” or heavily edit the output, perhaps to the point where creating it manually would have been more efficient.
It is this vague, ‘indeterminate space’, the realm of obfuscation and noise, what I like to call ‘the fluff’, that design thrives in. It’s in this space, the area that most businesses are too busy to deal with, or too deep in the weeds to make sense of, that we as designers must embrace, even more wholeheartedly, as a form of insulation against the more ‘definitive’ and binary-oriented ‘generative bomb’ that is beginning to impact so many other fields.
This insight, although not groundbreaking (I’m obviously not that quick), has alleviated the first of my concerns: “Does the design industry have a future moving forward?” “Hell, yes!” However, it didn’t diminish my second concern, my own personal job security, specifically my role as a Creative Director. This is because the “that’s good enough” mentality mentioned above led me to believe that most clients would, in the name of efficiency and cost, be satisfied getting 55% of the way there with their own in-house generative tools. That the creative ‘oomph’ required to take things from mediocre to magical would likely be viewed as an unnecessary expense, potentially marginalising creative specialists like myself who focus on driving teams and work beyond the mundane and take it to places that even I didn’t think possible.
I started picturing this future world filled with derivative, generative sludge, churned out by these ‘good enough’ clients. A rather depressing thought, I must admit. But then, a glimmer of hope! Perhaps this devolution presents an opportunity. Clarity, tone, cohesion and vision become the rare gems, almost non-fungible, even priceless. It’ll take a while, of course, for businesses to realise the pitfalls of DIY Creative Direction, but they’ll eventually self-correct as a result of muddled messaging and becoming lost in the noise… and, let’s be honest, did I really want to be working with clients content with generative mediocrity?
No!
My best work, the stuff that truly sings, comes from collaborating with clients and designers who aspire to create something exceptional. So, my initial negative fears have now morphed into a renewed sense of purpose and direction. I’m eager to work with those ever rarer clients who possess the vision and ambition to create something truly extraordinary. AI? It’ll be a tool, one of many, a tactical asset for some projects, a strategic partner for others. I’ll use it to enhance my processes, explore new avenues, but never to replace the human element – the depth, taste, nuance, and holistic understanding of good Creative Direction that breathes life into great brands. It’s this realisation, that AI is a tool to be mastered, not a force to be feared, that has illuminated my personal path forward.
So the future of design lies not in fearing the generative bomb, but in learning to love it, to harness its potential while staying true to design’s affinity with the 'fluff'. It's about finding that delicate balance, and remembering that even in a world of rapid technological advancement, the human touch, the ability to imbue work with emotional resonance, context, cultural understanding, and a unique perspective, will always remain invaluable, unreplicatable and profoundly marketable.
As such we as designers shouldn’t adapt to the storm; we should instead learn to conduct its lightning.