Written by:
Victoria Kulaeva
There’s something dangerously convincing about AI.
When we first start using it, we question everything. We check the sources, rewrite the copy, challenge the recommendations.
But then AI gets better. And we get more comfortable.
We stop checking every answer. We trust the research summary. Maybe we don’t even reread that piece of copy before pasting it.
And at some point, we’re no longer just delegating the task. We’re starting to delegate part of the thinking.
As a UX/UI Design Lead, I find this phenomenon much more interesting than whether AI can generate a screen or build a prototype. We already know it can.
So the question becomes: how do we use all of that capability to become stronger thinkers ourselves?
Using AI doesn’t automatically make you AI-literate
I want to be clear: I’m pretty optimistic about AI for work.
I use it every day, and I think designers who learn to work with it properly become far stronger at their craft. Figma’s 2026 State of the Designer research backs this up with 89% of designers saying AI helps them work faster and 91% saying AI tools improve their designs.
But faster output doesn’t remove the need for expertise. If anything, I think it raises the bar. When AI can generate 20 plausible directions in the time it once took us to explore two, generating options is no longer the hardest part.
Choosing is.
Which idea actually solves the problem? Which one makes sense for this user, this business, this product? What looks convincing but falls apart once you start questioning it?
Figma describes this shift well in The AI Creative Process.
That’s why I see a big difference between knowing how to use AI and being AI-literate.
Writing a great prompt is useful. Being able to recognize when the answer shouldn’t be trusted is much more important.
Treat AI like a brilliant, very junior colleague
This is probably my favorite mental model for working with AI.
Imagine a junior colleague who is swift, has read an absurd amount of information, never gets tired, and is willing to brainstorm with you at 2 a.m.
Sounds amazing, right?
There’s just one problem: they’re also keen to please you. As a result, they give you beautifully structured, confident answers without the deep context that you have.
So my rule is simple: TRUST, but VERIFY.
If I’m using AI for research and it gives me a statistic that could influence a product decision, I want the source. And then I actually open it.
- Does the source exist?
- Is it credible?
- Is it current?
- Who was studied?
- Does the research really say what AI claims it says?
To me, good research is close to scientific work. If we’re going to base a decision on something, we’re responsible for knowing whether that information is reliable.
“AI told me” is not a source.
And “AI made a mistake” isn’t really an excuse either. If I use AI in my work, I’m still responsible for the result.
Being able to build an app is not the same as being able to design a good one
I’ve looked at applications built by very technically capable people that, at first glance, are genuinely impressive. They look polished. There’s a lot of functionality. Everything seems to work.
Then, I start using them. And quite quickly, I start finding things:
- A flow takes five steps when it could take two.
- An interaction behaves differently from what users expect.
- Important edge cases weren’t considered.
It’s technically functional, but the experience of using it is far from easy.
A working interface built from a technical background and AI tools only isn’t the same as a good experience built with good design thinking.
Good UX is the result of hundreds of small decisions:
- what we show or don’t
- what we prioritise
- what users expect to happen next
- where they might hesitate
- how the system behaves when something goes wrong
- how all of that connects back to the business goal.
Interestingly, Google Senior UX Designer Katie Jacquez recently wrote about almost exactly this problem. AI can produce a polished, high-fidelity product remarkably quickly, but visual completeness can create a false sense of security. Behind a convincing interface, the interaction model or even the fundamental product logic can still be wrong.
It takes design experience to see what the polish is hiding.
This is why we invest in AI literacy
At Snappymob, our answer isn’t to protect ourselves from AI. Quite the opposite — we want our people to be very good at using it.
We learn better prompting and challenging methods, discuss new tools and workflows, pay attention to how sensitive information is handled, and share what works across the team. Recently, we also had an AI Literacy masterclass with Mustapha Benkalfate, an AI trainer and researcher whose work includes the cognitive risks of generative AI.
The objective isn’t to delegate more of our brains to machines.
It’s to use these tools to make us stronger specialists.
If I had to choose today between two designers — one with excellent UX fundamentals and strong critical thinking who barely uses AI, and another who knows every new AI tool but lacks those fundamentals — I would choose the first one.
And then I’d teach them AI. Because tools can be learned surprisingly quickly. Judgment takes years.
And perhaps that’s the part of the AI conversation we should be paying more attention to.
AI gives us access to an extraordinary amount of knowledge. But great designers bring something else to the work too. They bring not only “ratio” (reason, rational thought), but also “sensus” (feeling, perception). To me, good design has always needed both.
You need the knowledge and experience to understand a problem, the critical thinking to question your own conclusions, and the rationality to make good decisions. But you also need empathy. You need to notice people, listen to them, feel when something isn’t right, and genuinely care about the experience you’re creating for them.
I believe a good heart matters in design too. When empathy, sensitivity, and genuine care for people are combined with years of experience, knowledge, and strong critical thinking, you get something that is very difficult to replicate with a tool, no matter how capable that tool becomes.
So yes, we should stay curious enough to keep learning, critical enough to question what we’re given, and ambitious enough to use AI to become better at what we do.
But we should also remain human enough to remember who we’re designing for.
The opportunity isn’t to let AI think instead of us — it’s to use AI so we have more capacity to think better, while protecting the very human qualities that make good design good.