When I set out to build this website, I thought I'd take a shortcut. I'd prompt my way to a finished product - describe the layout, the typography, the vibe - and let AI do the rest. It would be fast. It would be easy.
It wasn't.
What I got back was fine. Technically functional, visually passable - but it didn't feel like me. The spacing was off. The font choices were generic. The hierarchy didn't guide the eye the way I wanted it to. It looked like what AI thinks a designer's portfolio should look like, not what I actually wanted to communicate.
It wasn't until I pulled in my own Figma designs, defined the visual direction myself, and used AI as a collaborator rather than a replacement that the site started to come together. AI handled the code. I handled the taste.
That experience crystallised something I've been thinking about for a while now.
Anyone can generate a design. Not everyone can design.
We're living in a moment where the barrier to creating a UI has essentially collapsed. You can prompt your way to a landing page, a dashboard, a mobile app screen - in seconds. And honestly, the output is impressive. It's getting better every month.
But producing a layout isn't the same as designing one.
Design has never been about the artefact. It's about the decisions behind it. Is this accessible? Does the information architecture actually make sense? Has this been tested with real users? Is this font size and colour combination readable for someone with low vision?
These aren't questions that AI is asking. These are the questions that designers ask - and when we stop asking them, we're honestly no better than the AI itself.
The role hasn't changed - it's sharpened
Here's what I think people get wrong when they talk about AI replacing designers: they're confusing the craft with the output. AI can produce the output. What it can't do is hold the line on quality. It can't advocate for the user. It can't look at a screen and feel that something is off before being able to articulate why.
The designer's role is still what it's always been - tastemaker, user advocate, and quality gatekeeper. If anything, it's become more important. Because when the speed of production increases, so does the risk of shipping things that look right but aren't.
AI has given us the ability to move faster. That's genuinely powerful. But speed without judgment is just noise.
What happens when we move too fast
I work with a team that's very AI-forward, and I've seen the benefits firsthand - faster prototyping, quicker iteration, less time on repetitive tasks. But I've also seen what happens when the pendulum swings too far.
Designs that look polished on the surface but lack consistency when you zoom out. Choices that were made because the AI suggested them, not because someone interrogated whether they were right. Components that technically work but don't feel cohesive as part of a system.
None of this is because AI is bad. It's because the design judgment layer got skipped. And that's on us, not the tool.
The moments where I've seen the most impact are the ones where a designer steps in and asks: Wait - have we actually validated this? Does this work for the edge cases? Is this the right pattern for this context? Those are the moments that protect the user.
Embrace AI, but don't betray yourself
I'm not writing this to argue against AI. I built this very site with it. I use it daily. It makes me faster, and it makes the work better in a lot of ways.
But there's a difference between using AI to amplify your skills and using it to replace them. If we stop testing with users because AI can generate a "good enough" solution - we've lost something. If we stop thinking critically about accessibility because the output looks clean - we've lost something. If we stop developing our own taste because we're deferring to what the model suggests - we've already been replaced. Not by AI. By our own neglect.
The designers who will thrive in this era are the ones who embrace AI for the speed and capability it brings, while doubling down on the things it can't do - empathy, critical thinking, systems-level judgment, and the quiet instinct that tells you something isn't right before the data confirms it.
AI is meant to make us better
Not just faster. Better.
That means better at advocating for users. Better at holding quality standards. Better at spending our time on the decisions that actually matter, because the production work is handled.
The role of the designer hasn't shrunk. The craft hasn't been devalued. If anything, the world needs us more - because there's about to be a lot more design out there, and most of it won't have a human behind it asking the hard questions.
Make sure you're still asking them.