We're Still Teaching UX for a World That No Longer Exists
Most UX curriculum was written for a world where the primary artifact is a screen. That world is changing faster than the training is.
The gap is widening
There's a gap forming in UX education. It is not a new gap. The field has always had some distance between what practitioners do and what programs teach. But that distance is widening, and it is widening in a direction that matters.
Most UX curriculum was built for a world where the primary artifact is a screen and the primary interaction is a person clicking through it. Journey maps assume linear flows. Information architecture organizes navigable structures. Usability testing evaluates how people move through interfaces. The whole apparatus is tuned to a specific model of human-computer interaction: intentional, sequential, direct.
That model still holds in plenty of contexts. But it is being supplemented, and in some domains replaced, by systems it was never designed to handle. The curriculum has not caught up.
The question has changed
The emerging design challenge is not "how does someone navigate this?" It is "how does someone trust this, oversee it, and intervene when it gets things wrong?"
Those are different questions. They demand different skills, different instincts, and different ways of framing a problem. And they are largely absent from how design education is structured, both in academic programs and in the professional development world built around the field.
Trust in AI-mediated systems is not a UI problem
Designing for trust in an AI-mediated system is not primarily a UI problem. What information does someone need to feel confident delegating a decision to a system? How do you communicate uncertainty without undermining confidence in the result? What does a meaningful intervention point look like when the system does most of the work and the human supervises rather than executes?
These are design questions. They belong to the discipline. Most curriculum does not treat them that way because, when the curriculum was written, these systems did not exist at scale.
Tool fluency is there. Vocabulary is not.
From what is visible in portfolio work and hiring conversations, candidates are arriving with strong tool fluency. Figma, research synthesis, component systems. The table stakes of the contemporary portfolio are largely there.
What is weaker is the conceptual vocabulary for the problems that are becoming more common. Not because candidates are not capable. Because nobody has asked them to think this way yet.
Ask someone to define a mental model and they will give you a solid answer. Ask them to design for a situation where the user does not fully understand the model the system is operating on, where the logic is probabilistic and the output is right most of the time but not always, and the frameworks that usually help start to wobble. I have watched this happen in portfolio reviews: smart designers, good instincts, and genuine uncertainty about how to approach a problem they have never been given language for.
That is not a failing of the candidates. It is a gap in what the field is pointing people toward.
"Teach designers about AI" is not the point
"Teach designers about AI" is not the takeaway here. I want to be specific, because it is where a lot of teams land and it mostly misses the point.
The skills that matter for this next phase are not primarily technical. They are conceptual, and many of them have been sitting in adjacent disciplines for years without being pulled into mainstream design education.
How does a system that makes decisions on someone's behalf change what that person knows, expects, and does next? That is systems thinking, applied to behavior over time, not just information architecture. What builds and sustains confidence in a system whose logic you cannot fully see? That is trust architecture: transparency, reliability, and knowing when to surface uncertainty rather than hide it. How do you keep a human meaningfully in the loop when most execution is automated? What does "wrong" even mean in a probabilistic system? How do you design for a distribution of outcomes instead of a binary?
These questions have homes in behavioral economics, human factors, and trust and safety research. Designers do not need to become experts in those fields. But designers do need enough grounding to work at the intersection without defaulting to UI solutions for problems that are not primarily UI problems.
Why education lags
The curriculum problem is a lagging-indicator problem. Programs teach what the field was doing five years ago because that is when the curriculum was written, and curriculum development is slow. That is not new.
What is new is that the shift happening now is not incremental. It is categorical. The conceptual frame, what a designed experience is, who is in the loop, and what interaction even means, needs updating, not patching. Adding an AI module to an existing program is like adding a chapter on GPS to a navigation textbook that assumes everyone is reading a paper map. The tool is different, but so is the model of how you get somewhere.
What to examine now
If you run a design team, the rubric you use to evaluate candidates and work probably bakes in outdated assumptions. It is worth examining which skills you are optimizing for, and whether they map to the world the field is moving into or the world it came from.
Journey maps are still useful. IA still matters. Usability testing is still how you catch certain classes of problems. But if those are the only lenses your team uses, you will miss issues that do not surface until after launch. I have seen that happen often enough that it stopped surprising me.
If you work in education, and I mean academic programs and corporate L&D alike, the "interface" assumption is baked into almost everything: what you teach, what you assign, how you grade. Surfacing that assumption explicitly, and questioning it rather than inheriting it, is where this starts. That is not a curriculum overhaul. It is a conversation worth having.
The gap is not new. But the cost of leaving it open is higher than it has ever been.
The skills the field needs next are knowable. Build toward them.