AI in practice

What I’ve seen less of, though, are architects writing about it, and specifically how they use it. Maybe because most of us are still figuring out how it fits into our workflow — whether it’s the threat we’ve been told it is, and if we do embrace it, how we can possibly explain that to clients without sounding like we’ve automated our own creativity.

So, I’m sticking my head slightly over the parapet here. I’m not going to prophesize about where the tech is taking our industry, or who stands to lose or gain. I’m simply reporting a little on how it’s influencing the way we work at Dune.

AI is everywhere in architecture right now — and while it’s undeniably raising the average level of visual quality, it’s also flattening the field. The result is what feels like a sea of sameness. Everything looks polished, perfectly lit, beautifully composed — but it’s harder than ever to tell what’s genuinely good.

Real-time rendering changed the game first. It let us test ideas live — moving materials, lighting, and proportions in real time. It made us bolder, more experimental. We could try things we’d never risk before because the feedback was instant and tangible.

Does AI do this better? Not yet. The inaccuracies are still too distracting to use for anything remotely presentational. It’s great for early massing studies, mood cues, and atmosphere, running several moodboard options, trying an option in red (just because you can) — but when you’re developing a real building, it falls apart on precision.

Are we embarrassed to use it? Not at all. Even as students long before AI, our unit leader encouraged us to use every tool available to best communicate our designs. Every year, a new piece of software arrived that halved rendering times or doubled realism. Was that cheating? No. It was progress — and this is simply the next step in that continuum. We spent less time waiting for the computer to do it's thing and more time designing and testing.

As AI becomes better at creating, expectations have risen too. People notice the small details now — the joinery scale, the soffit shadow, the way materials meet. When everything looks good, refinement becomes the real differentiator.

At its core, architecture is about decision-making. Architects use their experience to build a value system of priorities — balancing people, context, cost, and time. Unless AI has an intimate relationship with the client — their quirks, finances, and long-term goals — it can only make pictures, not purpose.

I like to think of AI as a helper, not a creator. It enables quicker decisions, allows faster exploration, and helps refine ideas. Ultimately, that means less time, less cost, and better clarity for our clients.

We’re moving at the speed of technology — but the craft of making remains, reassuringly, human.