Short answer: no, but the entry-level version of it is shrinking while the senior version is expanding. The job title is not disappearing from official employment data, and the tasks inside it are being redistributed. US Bureau of Labor Statistics data separates the two halves cleanly: web developers and digital designers are projected to grow 5 percent between 2025 and 2035, adding 11,300 jobs from a base of 220,100, while graphic designers are projected to decline 2 percent over the same period. Visual production is contracting. Interface and product thinking is not. That gap explains almost everything people are experiencing.
The Feeling Versus the Data
Why the field feels like it is collapsing
Three things happened at once, and they get blamed on a single cause.
Hiring normalised after an unusual boom. Design teams expanded rapidly during a period of cheap capital and then contracted. A junior designer entering the market now is competing against people laid off from those teams, which feels like the field dying rather than a cohort effect.
The portfolio-only route stopped working. For several years, a bootcamp certificate plus three case studies was enough for a first role. It no longer is, because there are far more candidates with that exact package than there are junior openings.
Tooling absorbed the visible part of the job. Generating layouts, variants, copy and states is now fast. Since that output was the most visible artefact of design work, its automation reads as the automation of design itself.
There is a fourth factor that rarely gets named. Design’s seat at the table was partly won by arguing that good interfaces were hard to produce. When production became cheap, that argument stopped working, and teams that had never fully understood what designers contributed beyond artefacts drew the obvious wrong conclusion. Rebuilding that case now requires talking about decisions and outcomes rather than craft, which is an uncomfortable shift for people who trained on craft.
None of these mean the discipline is dying. They mean the same amount of design work is now distributed across fewer, more senior roles, with a higher bar for entry.
What the employment data actually shows
The broader labour context matters here. BLS projects total US employment to grow 3.5 percent from 2025 to 2035, noticeably slower than the 10.9 percent of the previous decade. Against that baseline, a 5 percent projection for web developers and digital designers is above-average growth, not decline.
The contrast with graphic designers is the useful signal. Both roles involve visual craft. One is projected to grow and one to shrink. The difference is that digital design work is tied to systems that need maintaining, testing and iterating, while a large share of graphic design work was discrete artefact production, which is exactly what generative tools do cheaply.
BLS also published a new set of AI exposure categories alongside the 2025-35 projections, classifying every detailed occupation into Low, Moderate, High or Very high relative exposure by combining five external datasets. The agency is explicit that exposure “does not imply job loss, productivity gains, automation probability, or wage effects.” That caveat is worth holding onto, because most of the alarming coverage you will read treats exposure and replacement as the same thing.
What AI Genuinely Takes Over in UX Work
Being specific here is more useful than arguing about the field in the abstract.
| Work | What changed | What is left for a designer |
|---|---|---|
| Wireframes and layout variants | Generated in seconds | Choosing which variant to test and why |
| UI copy drafting | Drafted instantly | Voice, legal constraints, edge-case wording |
| Design-system component states | Auto-generated | System architecture and naming decisions |
| Research synthesis | Clustered automatically | Deciding what the clusters mean for the roadmap |
| Usability test transcription | Fully automated | Spotting what the participant did not say |
| Redlines and handoff specs | Largely automated | Negotiating the trade-offs behind the spec |
Two rows in that table deserve more attention than they usually get. Research synthesis is the one people most readily hand over, and it is the one where handing over costs the most. Clustering is genuinely useful, but the insight in qualitative research usually sits in the moment a participant contradicts themselves, hesitates, or works around a feature instead of using it. A clustering pass flattens exactly those signals into a tidy theme.
Usability transcription is the opposite case. It was pure overhead, nobody enjoyed it, and automating it returns hours to the part of the job that actually needs a person. A designer who spends the reclaimed time watching sessions again rather than filling it with more tickets comes out ahead.
The pattern is consistent. Production is compressed. Judgement is not. A designer whose value came from producing artefacts quickly has lost most of that value. A designer whose value came from deciding what to build, and defending it against a product manager and an engineer with different priorities, has lost almost none.
What to Know Before You Panic or Pivot
Portfolio inflation is real and it is not about you. Reviewers now see many portfolios containing beautifully generated screens with no evidence of decisions. The differentiator has moved from visual quality to demonstrated reasoning: what you tested, what surprised you, what you changed.
Job titles are drifting. Roles that would have been posted as “UX Designer” now appear as product designer, design technologist, or design systems engineer. Searching only for the old title will make the market look emptier than it is.
Generalists are being squeezed from both ends. Specialists in research, systems, or complex regulated domains are still in demand. So are designers who can build. The middle, the person who makes competent screens from clear requirements, is the most exposed position.
Salary bands are separating. The gap between a designer who can only produce and one who can decide and build is widening in compensation as well as in hiring speed. That divergence is a better guide to where to invest your time than any prediction about the field.
The best evidence is boring evidence. A single case study showing a measured behavioural change beats five polished redesigns of well-known apps.
The Skills That Now Separate Hired From Not Hired
- Framing problems, not just solving them. Being able to argue that the brief is wrong, with evidence, is the most senior skill in the discipline and the least automatable.
- Working prototypes. Designers who ship interactive prototypes in code, or with code-generating tools, remove an entire handoff cycle. This is currently the strongest single differentiator in hiring.
- Design systems thinking. Naming, tokens, governance and versioning are architecture problems. They stay human because they are political as much as technical.
- Measurement literacy. Reading an experiment result, understanding when a result is noise, and knowing what a metric will do to behaviour.
- Directing AI tools deliberately. Not “I use AI” but “I use it for divergence and never for the final decision, and here is the audit trail.”
- Domain depth. Healthcare, finance, industrial software and accessibility all have constraints that a general-purpose model handles badly.
A Decision Framework for Your Next Two Years
Work out which of four positions you are actually in, because the right move differs sharply.
- Trying to enter the field. Do not compete on portfolio polish. Pick one narrow domain, build one real thing for real users, and measure something. One documented outcome beats six speculative redesigns.
- One to three years in. Your risk is being the fastest producer on the team. Move deliberately toward research, systems or prototyping in code within the next year.
- Senior with no AI workflow. Your risk is credibility rather than capability. Rebuild one recent project using current tooling and document what improved and what got worse.
- Considering leaving design entirely. Check whether you dislike design or dislike your current employer’s version of it. Product management, research and design engineering all reuse most of your existing skill set.
A worked example of the second case makes it concrete. A designer three years into a product role spends most of her week producing screens from written requirements. Her output is good and fast, which is precisely the risk. Over six months she moves two days a week into running lightweight tests on the features she is already designing, starts writing the test plan before the screens, and ships one prototype in code instead of a static file. Nothing about her title changes, but the story she can tell in an interview changes completely, because she now has evidence that her decisions moved something.
For every path, the same practical test applies: can you show a piece of work where a decision changed because of something you found out? If yes, you are employable. If no, that is the gap to close first, and it has nothing to do with AI.
Building AI Fluency Without Leaving Design
The designers doing well right now are not the ones who resisted the tooling or the ones who handed everything to it. They are the ones who worked out, task by task, where the model helps and where it quietly degrades quality, and who can explain that split to a hiring panel.
That fluency is learnable in weeks rather than years, but it does need structure, because experimenting at random produces habits rather than judgement. A guided sequence plus one real project you rebuild with the tools gives you something concrete to talk about, and pairing that with a certificate makes it legible on a CV as well as in an interview. If you want a structured starting point, explore Coursiv AI lessons and check current plan details on the official site.
FAQ
Is UX design still a viable career?
Will AI replace UX designers entirely?
What should I learn first if I am worried?
Which parts of the job are safest?
Are bootcamp graduates still getting hired?
What to Do This Month
Take your strongest recent project and rewrite the case study around decisions instead of screens: what you believed, what you tested, what changed your mind, and what happened to the metric. Then rebuild one screen of it with current AI tooling and note precisely where the output needed correcting. Those two artefacts answer the question a hiring manager is actually asking, which is not whether you can design but whether you can think in a market where designing got cheap.