■ HIGH RISK ■ Creative Arts
AI now produces competent screens, flows, and even research synthesis on demand, which guts the production-design middle of the field. Designers who own strategy, research judgment, and product decisions survive; designers who own Figma files don't.
“AI A/B tests 1,000 layouts before you've opened Figma.”
Our AI replacement risk score — how we score jobs
UX design's day-to-day spans a wide arc: user research and interviews, journey mapping, wireframes and prototypes, high-fidelity screens in Figma, design-system maintenance, usability testing, and the endless meetings where design negotiates with product and engineering about what actually ships. In big companies the role fragments into specialties; in startups one designer does all of it before lunch.
AI has arrived at nearly every stop on that arc. Text-to-design tools generate credible screens and full flows from a prompt; Figma's own AI features draft layouts, rename layers, and produce variants; code-generation tools like the current wave of prompt-to-app builders let PMs ship passable interfaces with no designer involved — that last one is the real threat, since it removes the request rather than speeding the response. Research is squeezed too: AI transcribes, tags, and synthesizes usability sessions, and synthetic-user testing claims to simulate feedback panels. Production tasks that filled junior designer days — spacing tweaks, component variants, responsive adaptations — are increasingly a model's output awaiting review.
What resists is everything upstream of the artboard. Framing the right problem, choosing what not to build, research with actual humans whose behavior contradicts their words, navigating stakeholder politics, accessibility and ethical judgment, and crafting novel interaction patterns rather than remixing existing ones — AI generates from the corpus of what exists, which is precisely why AI-generated interfaces converge on the same competent sameness. Demand dynamics cut both ways: cheaper design means more products get designed, but hiring data since 2023 shows fewer junior openings and rising expectations per designer. Our 61 reflects a profession whose output is commoditizing faster than its judgment: fewer, more senior, more product-strategic roles.
Automatability: our editorial assessment of current and near-term AI capability
Underway now, hardest at the entry level. Screen production and research synthesis are automating today, and prompt-to-app tools are removing some design requests altogether — junior UX hiring already reflects it, and the pressure compounds through the late 2020s. By 2030 the field consolidates around senior product-minded designers directing AI output; the traditional apprenticeship path of learning by pushing pixels will be mostly gone, which is the profession's quiet succession crisis.
With clear eyes, maybe. The junior market is genuinely hard — the production tasks that entry-level designers were hired for are the most automated part of the field. If you enter, come with adjacent strengths: research rigor, front-end literacy, or domain expertise, and expect to use AI tools from day one. The 'learn Figma, get hired' era is over.
It does the visible parts convincingly — generating screens, flows, and variants that look professional. What it doesn't do is decide whether those screens solve the right problem, reconcile conflicting user evidence, or push back on a stakeholder's bad idea. The output is design-shaped; the judgment isn't there. That gap is exactly where the remaining jobs sit.
Problem framing, research with real users, strategy, facilitation, and accessibility/ethical judgment — plus the political skill of getting good decisions made in organizations. Visual production is the most exposed; research synthesis is partially exposed. The consistent pattern: work that touches humans and decisions endures, work that touches artboards automates.
Learn to ship, which is subtly different. Prompt-to-app tools mean designers can now build working products without deep engineering skills — and designers who do so become product creators rather than deliverable producers. Understanding front-end fundamentals still helps you direct AI output credibly. The valuable position is 'person who turns intent into shipped product,' whatever mix of code and prompts that takes.