This is not a "UX is dying" story
AIPERSONAL DEVELOPMENTBRANDING
For years, conversations about AI and design have revolved around one anxious question:
Will AI replace UX designers?
It is understandable, but it may be the wrong question.
A UX role is not one indivisible unit of work. It is a bundle of activities: analysing evidence, writing copy, producing wireframes, facilitating decisions, reframing problems, negotiating constraints and understanding people.
AI will not affect all of those activities equally.
That is what makes MIT’s Iceberg Index useful. Instead of treating an occupation as either “safe” or “at risk,” it examines where current AI capabilities overlap with the skills used inside an occupation.
The result is not a prediction of which jobs will disappear. It is a map of where the work is already beginning to change.
The important distinction: skills, not jobs
A job title hides enormous variation.
Two people called “UX designers” might spend their weeks doing completely different things. One may focus on research and strategy, while the other produces interface designs inside a mature design system. A third might spend most of their time aligning stakeholders rather than creating screens.
Job-level predictions flatten those differences.
The Iceberg Index takes a skill-centred approach. Its score represents the share of wage value connected to skills where current AI systems demonstrate technical capability.
That wording matters.
According to the project’s official FAQ, a high score does not mean the same percentage of jobs will disappear. The index does not predict layoffs, adoption timelines or workforce reductions. It measures technical overlap.
Actual automation depends on factors the index does not attempt to model: organisational readiness, cost, regulation, trust, incentives and whether people are willing to delegate the work.
In other words, the Iceberg Index is a diagnostic tool, not a crystal ball.
What happens when we apply this lens to UX?
I mapped the framework to six common areas of UX work:
Data analysis and metrics: 80%
UX writing and microcopy: 75%
Competitive analysis: 72%
Wireframing: 68%
Usability testing: 50%
Strategic design thinking: 22%
These figures are my application of the framework to UX skills, not scores published by MIT for the UX profession.
They should be interpreted as directional signals rather than precise forecasts. Their value comes from the pattern they reveal.
The tasks with the highest overlap are largely tasks that produce visible outputs:
A summary of analytics
A set of interface messages
A competitor comparison
A wireframe
A usability report
These are also the artefacts designers often use to demonstrate that work has happened.
That creates an uncomfortable possibility: the parts of UX that are easiest to show may also be the parts becoming easiest to reproduce.
The deliverable was never the real value
A wireframe is visible. The reasoning that shaped it usually is not.
A research report is visible. Knowing which questions to investigate, which behaviours matter and which findings deserve action is harder to see.
Microcopy is visible. Understanding the emotional state of a person encountering it, the commercial pressure surrounding it and the legal constraints behind it is mostly invisible.
For a long time, this distinction did not seem urgent. Producing the artefact required enough specialist knowledge that the output became a reasonable proxy for expertise.
Generative AI is breaking that relationship.
A polished deliverable can now appear before the underlying problem has been properly understood. A plausible answer can arrive without anyone deciding whether the right question was asked.
That does not make the deliverable useless. It makes it insufficient evidence of value.
The future value of UX will depend less on our ability to produce an artefact and more on our ability to explain why it should exist, what informed it and which trade-offs it resolves.
What sits below the waterline?
The less exposed parts of UX are not mysterious creative superpowers. They are demanding human activities that happen inside a specific context.
Problem framing
AI is highly capable of generating solutions to a stated problem.
The difficulty is determining whether the stated problem is the real one.
A request to improve conversion might actually be a trust problem. A request for a new dashboard might reflect unclear ownership. A request to simplify a form might be constrained by regulation rather than poor interaction design.
Good framing changes what the team builds and sometimes reveals that it should not build anything.
Judgement
Design decisions rarely have one objectively correct answer.
They involve incomplete evidence, competing goals and consequences that cannot always be reduced to a metric. Designers decide which signals matter, which compromises are acceptable and when more certainty is not worth the cost.
AI can contribute options and analysis. Accountability for the decision remains human.
Stakeholder trust
Products are made through groups of people with different incentives.
A designer’s impact often comes from creating enough shared understanding for those people to move together. That requires listening, negotiation, timing, empathy and the ability to challenge someone without losing their trust.
These capabilities are difficult to see in a portfolio, but they determine whether good ideas survive contact with an organisation.
Emotional and cultural understanding
People do not experience products as neutral task-completion systems.
They arrive worried, distracted, excited, sceptical or under pressure. Their interpretation is shaped by language, identity, culture and previous experiences with the organisation.
AI can identify patterns and simulate perspectives. It does not participate in the consequences of getting those interpretations wrong.
UX is not dying. Its centre of gravity is moving.
The shift is not from “human designer” to “AI designer.”
It is from production-centred UX to judgement-centred UX.
Designers will still produce flows, interfaces, prototypes and research outputs. But more of that production will be accelerated by AI. The differentiating value will increasingly sit in deciding:
What deserves attention
Which evidence should be trusted
What problem is actually being solved
Who needs to be involved
Which trade-offs are responsible
How a proposed change will affect real people
Whether the output is good, merely plausible or fundamentally misguided
This is not necessarily bad news for UX.
It may force the profession to become clearer about the value it has always claimed to offer.
What should designers do now?
The answer is not to reject AI or compete with it on production speed.
That is a race the tool is designed to win.
Instead, designers can use AI to reduce the time spent creating first drafts, synthesising routine information and exploring obvious alternatives. That time can be redirected toward the work that is frequently squeezed out of delivery cycles.
Three changes are especially important.
1. Make your reasoning visible
Do not present only the final screen or document.
Show the assumptions you tested, the options you rejected, the evidence that changed your direction and the trade-offs behind the result.
2. Describe outcomes beyond deliverables
“Created a prototype” explains what you made.
“Aligned three teams around a shared definition of the problem and prevented an unnecessary build” explains your value.
3. Treat AI capability as a changing design constraint
The Iceberg Index reflects current technical capability. It will evolve as models, tools and adoption patterns change.
The goal is not to find one permanently safe skill. It is to develop the ability to recognise change early and adapt your practice deliberately.
Explore the UX skills yourself
I created an interactive version of this analysis so you can explore how the MIT framework maps to different areas of UX work.
Explore the interactive UX Iceberg Index below
The useful question is not simply whether AI will change UX.
It already is.
The more important question is whether we continue defining our value through the work above the waterline or learn to make the deeper work visible.
Join the conversation
Visit my LinkedIn post:
Follow me
Idea's that drive innovation through curiosity
© 2025. All rights reserved. | Copyright and Terms of Use Agreement
Resources
- Blog
- Free Tools