Who Designed That?

The user tapped a button that hadn't been there the day before. The interface had generated it based on context - a shortcut to a task they do every week, surfaced at the moment they'd normally go looking for it. It was useful. Genuinely useful. And somewhere in the room, someone asked the question that was about to change how everyone there thought about their jobs.

"Who designed that?"

Nobody had a clean answer. An engineer had written the model. A product manager had defined the signals it used. A designer had approved the component library the interface drew from. But the decision to surface that button, in that context, at that moment - nobody had made it. The system had.

That question is going to keep coming up. And the answer matters more than it might appear, because the answer determines accountability.

Generative UI is arriving. The specific form varies: AI that selects which components to surface, systems that generate personalized layouts, interfaces that adapt in real time based on what a user needs right now. The particular mechanisms are less important than the pattern. AI is beginning to make decisions that used to belong to designers. The interesting question is what that means for design - and the interesting answer is not what most people assume.

The natural assumption is that AI generating UI means less design work. That assumption inverts the actual dynamic.

When a human designer decides what to show a user, the accountability is clear and local. One person made one call. When AI generates experiences dynamically, that AI is operating inside a framework that someone defined. The rules it follows, the constraints it respects, the values it optimizes for - someone wrote those. That someone is doing design work. And it is the most consequential design work in the product.

Designers who define the frameworks governing AI decisions are working at a different level of abstraction than designers who move a button four pixels to the right. They're not deciding what this particular user sees in this particular moment. They're deciding what every user of this system can see, under what conditions, shaped by what values. A framework that prioritizes clarity over options creates different experiences than one that prioritizes control. Those choices are design choices. The users of those systems live inside them.

What this asks of designers is different, but not foreign. It asks for systems thinking alongside object thinking. The ability to define principles, not just apply them. Judgment about what rules to set, not just which component to pick. These are skills designers already use - they show up in design systems, interaction principles, content strategy, accessibility standards. The shift is that those outputs move from guidelines that might or might not be followed to rules that shape every AI-generated experience downstream.

Other functions are involved in this work too. Engineers build the systems. Product managers define success metrics. None of them do this poorly. They work with the tools their training gave them. What designers specifically bring - comfort with ambiguity, judgment about human experience, a practiced sense of when structure serves people and when it constrains them - is exactly what framework-setting requires. Not because designers are the only people who care about users. Because this particular job calls for skills design has been building for decades.

When designers are the ones defining those frameworks, users gain something that doesn't show up in a component library: someone with expertise in human experience was present when the rules were made. Not just an engineer optimizing for performance. Not just a product manager optimizing for engagement. Someone whose career has been spent thinking about what it feels like to be a person trying to accomplish something.

"Who designed this?" is going to be asked more often. The designers who make it possible to answer that question clearly - because they were the ones who defined the rules - are doing something that matters. Not for design as a profession. For the people on the other end of those AI-generated decisions.