We’ll all have them, we get through the hype and see what the current breed of AI tools can do for us. The weighing up of the pros and cons, debating whether we’ll struggle to eat in future or be wide-eyed with the possibilities.

In the design systems space, we often struggle with communicating the value…of the idea of a system in itself to the value that our teams provide, which is where my concern lies. The part of my intro I mull over is what “AI tools can do for us”. My further pessimism is that it’d be short-sighted to think we personally gain from this and that the gain others might find would be at our expense.

There are loads of potential applications for AI tools in design systems — loads of them! Train an LLM on your documentation, and it could apply your components through a prompt! Powerful stuff…so why would any designers and engineers be needed?

These tools are damned impressive, having seen one debugging its own output being one of those moments of surprise I had recently. New tech like this doesn’t appear to be inherently good or bad, and we’ll find loads of positive outcomes as we ponder the bigger issues, but I feel like in design systems, we sit in a precarious place. If we can’t be persuasive and advocate for the value we provide, in a future where a form of AI basically is the system and can generate near production-ready output, what’s our role?

Skipping forwards, we might have agents using sites and services made by other agents in a world where we’re largely redundant. Those with experience of ‘the before times’ can correct and guide these agents, but how much would that matter?

Exploring the negative doesn’t mean there isn’t hope. These are tools, and we can define how we use them, to a degree. What we can do is get really good at this question of value and really go deep on that. What is it that we do that (so far) can’t be replicated? What can we provide that isn’t easy for an AI to riff off of prior works?

For product teams, how comfortable are you with output created by a form of AI from your design system that reduces your input in most tasks? How can we position ourselves beyond generating output and bring to the fore what we ask professionals and individuals really bring that to date machines can’t?

That’s the question more generally for humanity I guess but we can bring it to our doorsteps first through understanding our value, the types of value that our work brings and how we might explore uses of these tools that work in tandem with that over replacing us.