Let me be upfront: I use AI tools regularly. They’re part of my workflow, they’ve made certain parts of my job faster, and I’m genuinely curious about where the technology is going. I’m also not naive about what we’re trading for that convenience. This isn’t a hot take. It’s just where I actually stand.
The parts AI is genuinely good at
I use AI as a thinking partner more than anything else. When I’m stuck on a UX problem — an architecture decision, a component that needs to serve too many contexts, copy that isn’t landing — I’ll talk it through. Not because I expect a perfect answer, but because articulating a problem clearly enough to explain it to someone else is half the work of solving it.
I also use it for research scaffolding, pressure-testing my rationale before I commit to a direction, and on the development side of my workflow, for translating design decisions into clean, structured code. It’s fast. It’s useful. It handles a lot of the mechanical lift that used to eat up time I’d rather spend thinking.
For those things, I have no complaints.
The parts AI gets wrong — consistently
Here’s what I’ve learned from experience: AI does not have taste.
It can generate UI. It cannot tell you whether that UI feels right. It doesn’t know if your design system is starting to drift, if a spacing decision is slightly off in a way that makes the whole screen feel unsettled, or if a color combination is technically accessible but visually exhausting. Those judgments require a human eye that has spent years developing an intuition for what works and why.
Every time I’ve let AI drive visual decisions without staying closely involved, something has gone subtly wrong. The cohesion breaks down. Individual elements might be defensible in isolation but the whole stops feeling intentional. Design is a systems problem and AI, for all its capability, does not yet understand systems the way a practiced designer does.
That’s not a knock on the technology. It’s just an honest assessment of where the ceiling is right now.
The bigger picture I can’t ignore
I’d be leaving something important out if I didn’t say this: I think about the environmental cost of AI more than the industry talks about. The energy consumption of large language models is significant and growing, and “it’s useful” isn’t a complete answer to that.
I also think about the people. There are real conversations happening in boardrooms right now about replacing subject matter experts — writers, designers, researchers — with AI to cut costs. And I think those organizations are going to regret it. Not because AI isn’t capable, but because the value a senior designer brings isn’t just output. It’s judgment, context, advocacy, and the ability to ask the right question before a single pixel gets placed. You cannot automate that. You can only pretend to until something goes wrong.
I don’t say this from a place of fear. I say it from a place of having watched what happens when process replaces thinking.
What balance actually looks like for me
I use AI as a collaborator, not a replacement for my own expertise. It handles scaffolding. I make the calls. It generates options. I decide what’s right. It writes the first draft of a function. I review it line by line.
The creative decisions — the ones that require taste, systems thinking, and an understanding of the human on the other end of the screen — those stay with me. That’s not protectionism. That’s just good design practice.
The designers I think will thrive in the next decade aren’t the ones who resist AI or the ones who outsource their judgment to it. They’re the ones who stay curious, stay critical, and stay in the driver’s seat.
That’s the balance I’m aiming for. I don’t think I’ve got it perfectly figured out. But I think asking the question honestly is a pretty good place to start.
