The AI to UI trap
The tool defines the work
Lately I’ve observed something interesting in the AI conversations happening around me. There seems to be a stark contrast between what is said and showcased online versus what is believed and practiced in reality.
The online narrative is clear. You either get on board fully or be left behind forever. Designers are rebranding themselves as 'builders' and 'AI-native'. Vibecoded products are showcased, workflows are optimised and anyone not visibly adapting AI is quietly implied to be falling behind. The pressure is real and it's loud.
However, when I talk with design friends and colleagues, I see a different perspective. The pull towards AI doesn’t feel as strong. If anything, it seems like there is reluctance nobody quite wants to say out loud. In these more honest conversations, we ask each other how much we actually use AI. We share our hesitation to go fully in on a hyped tool, knowing the crowd will move on to the next hyped tool in mere weeks. We admit that we don’t always know how to apply AI purposefully.
My impression is that designers as a group appear to be more hesitant to adopting AI than our engineering and PM counterparts. It’s easy to dismiss this sentiment by saying designers are simply slow adopters. Resistant to change. Protective of their craft for sentimental reasons. But I don’t believe that can be the only reason.
For a long time, designers made it a point that their work isn’t something that solely lives in the shipping phase of the product development cycle. We crawled our way out of the basement to the seat at the table, wanting to influence strategy and prove we do more than pretty pictures. We reminded stakeholders that they hired us to tell them to pause and reflect if and how business goals can be aligned with customer needs and satisfaction. This made us an easy scapegoat, because companies have always looked for reasons to compress the problem definition phase. It’s slow, expensive and inconvenient. AI hands them a good reason to skip this phase altogether. Why reflect when you can ship, learn, and iterate in half the time?
That logic makes sense on the surface. AI tools can generate UI and code faster than ever. So lean in, ship faster, explore more. The gap in this logic is that the exploration AI enables is almost entirely limited to the execution phase. It makes building faster. It does not make thinking better.
The tool defines the work
The challenge with adopting AI in our field is not that AI might end up doing our work. The challenge is that it redefines what our work is.
Technology doesn’t just help us do things, it reshapes how we see the world and our place in it. A hammer makes the world look like nails. A camera makes the world look like images waiting to be captured. And Gen AI tools? They make every design problem look like a visual output problem. It reframes the entire territory of design as: what should this screen look like? And if that’s the question the tool answers, it quietly becomes the question we ask.
When we orient our workflows around a tool’s capabilities, we don’t just change how we work. We change what we think design means. And if the tool can only generate UI then design -from that perspective- becomes only generation. This reframing is what threatens our field.
As Damian pointed out, trying to compete on an execution level with AI tools is a race to the bottom. We can never outperform an AI when it comes to generating output. And thus the irony is that the technology we’re urged to embrace is the same technology that shrinks our role and makes us replaceable. But only if we let that technology define our role.
Reluctance as professional instinct
When designers are hesitant to adapt AI tools in their workflow, it’s not because they fear a new technology or they’re tired of learning yet another design tool. I see something that looks a lot like professional instinct. The instinct that makes a good designer pause before a stakeholder who says “just ship it” is the same instinct that makes designers feel uneasy about the adoption of AI tools. It’s the instinct that asks “but what problem is this actually solving?”
There’s a concept in systems thinking called drift into failure. It describes how complex systems don’t collapse suddenly but instead they degrade gradually. A series of individually rational decisions that together move the system toward catastrophe.
This is what I think is happening in design right now.
Each designer who rebrands themselves as a ‘builder’ does so out of survival instinct. Each individual choice (from either a designer or a company) to adopt an AI tool is rational. It saves time. It looks productive. It earns approval in a business environment that rewards speed. No single designer is making a catastrophic choice. But as collective, we may be drifting toward a world where design means execution. The problem definition phase is squeezed out and reflective pause is replaced by the fast prototype.
In conclusion
When I say that low AI adoption among designers might be self-preservation, I mean something more than protecting our job. I mean preserving the idea that design is a thinking discipline, not an output discipline. We’re hired to ask questions where the answer isn’t just a UI. Part of our job is to stop and redirect, to question if we’re solving the right thing especially when under pressure to ship.
I know these things are harder to defend in a sprint review. ‘Holding space’ is something Ariana Grande says in a feverdream Wicked interview, not something to show up in velocity metrics. The reflective nature of our job is increasingly in tension with how modern product organizations are structured. But in my opinion they are the core of what design actually contributes: the part that cannot be generated, only practiced.
So maybe we should listen to the quiet resistance in how designers are responding to AI. It might be the most important signal in the room.
How do you look at AI adoption in your work?
If you’re a returning reader: thanks so much for being around. If you’re a new reader: hi! My name is Maureen and I work as a senior product designer and content creator based in Amsterdam. If you liked what you read, please consider subscribing. All my posts are freely available and don’t require a paid subscription.






Your article made me think and realise that I will not win the AI output race. I have been using AI tools, and my hesitation is not about the possibility of being replaced but the need to use them by force, otherwise I can be fired. This is the sentiment. But the hesitation also came when I perceived all the output as something "pasteurised". Like comparing UI with an orange juice, and when you want an orange juice, you want that fresh flavour, but then with AI tools, you receive a "Juice of box" flavour, a pasteurised one, everything looks like the same thing that any company can create. My hope is that, as occurs with the juice, some people will still accept only the fresh ones. My reflection does not mean that I can't see value in the AI tool output, to ideate, to test interaction, this is amazing, but I don't believe companies should ship this "pasteurised" solution to the final user as a success path.
Thoroughly enjoyed reading this and appreciate the way you shared your thoughts with tact and care. I think it’s become so easy these days for a topic like this to fuel the AI fire either way (whether someone is for or against) but it rarely speaks to the actual reasons behind people’s hesitation with using AI.
I do resonate with the designers who feel peer- or client-pressured into using the tools against their instincts or better judgement just for the sake of keeping up with the times but not being able to clearly communicate *why* exactly something feels off.
I am definitely all for tools that stretch my skills and creativity (this includes AI), but still knowing to my core what my value is outside of any tools I may use (many of which will probably be ‘in or out’ in some way at some point in time).
But your essay is exactly the kind of thinking that helps people like me to articulate that value louder and more clearly.
Yes to all of it!🙌🏼