Generative AI in design
Lofty as it might sound, before we get into any specific discussion of how AI changes or supports design, we need to set out the territory that this debate occupies, we have to map out the philosophical and moral landscape. I promise we’ll be quick.
Some light philosophy
It’s easiest to first set out the inarguable facts, fixed points in our thinking. First, AI is here now and it’s not going away. Clichéd pandoras box metaphors aside, the technological progress that has allowed the current generation of tools to exist cannot be reversed, nor can it be ignored, and barring the end of the world, it will only advance. So we have to decide what to do about it. Second, it will affect every discipline, ours included. We should try and explore this idea – mindful of how everything around design will change and become normalised rather than in isolation.
Lastly, if AI is irrepressible and its effects inevitable, we each have to make a choice as to how we respond to it. There will undoubtedly be regulation, it is unlikely to be universal, so it will be conceived and implemented most likely on a national level. There may be some consensus and alignment internationally or supra-nationally, and there may also be some consensus or normalisation at an industry or organisation level (the SAGAFRA action in the US is a great example), and this will create variation, but it will be national governments that will set the pace and tone of AI development and usage, once legislation catches up with technology. It only takes one government to relax regulation for everyone else to have to do the same or risk being outcompeted.
Creative Value
So if you’ve come with me on this so far, where this leaves us, int he current absence of regulation, with a set of choices to make, not just about what AI can be useful for, but what the role of a designer is in the act of designing something, maybe even the role of people more generally. What do we want to keep in the hands of people and do we even have agency to decide int he long run?
One thing that has been a consistent theme of conversations I have had over the last 5 years with my teams about their career development, and growth as designers, is that more sophisticated tools commoditise the more repetitive and mechanical aspects of what we do. This has been going on for a while. At the start of my career we worked in photoshop and we manually created red line guides for developers to implement our designs. Then some bright spark created a plugin that generated these guides automatically (as long as you had been diligent in organising and grouping your layers) and, thought they had to be checked and manually exported as images, my workflow changed overnight. So did the amount of time I had to think about he actual work, rather than documenting it. Then along came zeplin and not only was this aspect of the work removed entirely, so was the value we had created in doing it.
In this case undoubtedly for the better, but i think this starts to illustrate the role of more sophisticated tools in chipping away at the nature of what we do and how we add value. Very soon AI will be sophisticated enough to move further up the value chain. It way well be that AI does the first pass for example at a user journey, or a wireframe, perhaps based on a human input, but perhaps even just based on a prompt or given the raw data about the problem we’re trying to solve for. If we allow this thought experiment to come to its logical conclusion, almost every aspect of the design and even more broadly the creative process can be performed by AI. Where this might leave us, and where I suspect we’re heading, is that the human role in design will be more one of curation and direction, rather than one of undertaking what we consider to be much of the act of design today. AI will produce design concepts, wireframes, strategies, images, and the role of the designer will be to course correct, to steer, to apply taste to adjust and adapt what an AI creates to meet definitions of human taste or to input into the process the human intuition and contextual understanding. We will steer ideas along a path, nudging them into reality, at every stage selecting from possible paths the work could take, or steering them on to new ones, some of which we might not have thought of otherwise. We will curate, shepherd, guide ideas from concept to reality, but it will require less of us, it will require different skills, and over time, we may find we lose the ability to undertake these steps ourselves, but it might not even matter.
Backlash
This beautifully uncomplicated experiment, fun as it may be, doesn’t actually answer any of the immediate and more pressing questions we have. Namely – what should be embrace, what should we be wary of, and where should we focus our energy and attention as we all orient ourselves to this new reality.
In the face of the industrial revolution there were many counter-movements that sought to fight progress deemed destructive or that undermined human value. We may well see this again, but for now, we have to make judgements on a daily basis. We can’t reject AI arbitrarily out of some sense of fear or insult at the idea of human creativity being devalued, but also not simply accepting every possible innovation with open arms without considering its implications.
Conclusion
We should embrace AI tools that take us to places we wouldn’t have otherwise gone, we should use it to broaden our perspective, to deepen our understanding and the come up with ideas we otherwise couldn’t. We should be wary of things that seem like an answer and solution to the commercial demands on us to work harder, to get to the end result quicker – if they compromise quality for a perceived cheaper or quicker result. We should be wary of things that remove humans from an inherently human act, like creativity. If all we see around us in our public spaces or interact with in our digital lives are things that AI has decided is important and the right thing to do then we will be poorer for it.
We should focus our energy on things that remove the aspects of our discipline that are less valuable, that divert time and energy away from the more valuable aspects. Like me working late to manually draw red lines and pixel measurements on photoshop files. We should though, in doing this, keep one eye always on how we describe our value, and what we might need to preserve in order to keep the inherently human value of art and design, and to make sure it is not cheapened or degraded in service of commercial aims.
We’re in the process of mapping this out in detail, identifying tools that are worthy of experimentation, where use cases for AI are unambiguously helpful, and what the considerations are where it’s not so clear. I’d love to show it to you.
