Why this job stays backstage
Ask whether AI will replace costume attendants and the answer sits in the shape of the work. A costume attendant hands out wardrobe before a show, dresses performers, runs quick changes in the wings, and mends a torn seam between scenes. Those tasks happen in a dark space, on a clock, with a person who is already in motion. Software has nothing to grip and no body to kneel beside.
The second half of the job is care of the clothes themselves. Washing, pressing, steaming and storing costumes so they survive eight shows a week. Checking pieces in and out, and spotting a loose hook before it fails on stage. All of it is judgment through the hands: how much a fabric will take, whether a repair holds for one more night.
Paperwork is the softer edge. Inventory lists, run sheets, fitting notes, research into period dress. That is where AI shows up first, and that is the part of the day most likely to shrink. The honest story here is task erosion, not a job disappearing. You can see the split on this page, and see how the headline figure is built on the Still needs a human method page.
What AI handles, what it assists, what people keep
Record-keeping is the clearest target. Costume inventories, check-in and check-out logs, repair histories and purchase lists all live in spreadsheets, and language models handle that kind of text well. Design and research tasks are also partly covered: image tools can sketch a look or pull period references for a designer to react to. The share of task time AI could take on today is 0%.
Assisted work is the middle band. Pulling a costume plot together, drafting fitting notes, estimating fabric needs, or translating a designer’s brief into a shopping list goes faster with a model in the loop, but a person still signs off. Scheduling dressers across a run is another one: the software suggests, the wardrobe supervisor decides. Task time where AI helps rather than does: 23%.
Everything physical stays with people. Dressing performers, running a thirty-second change, pinning an alteration on a moving body, pressing a delicate fabric without scorching it. Task time that still needs a human: 77%. Our Can AI do it? figure for this job is 16 out of 100, and the coverage method page explains what that counts.
How strong is the evidence?
Thin, and we grade it that way. The Is it better than a person? evidence grade for costume attendants is D, which means no study has tested a machine against a trained wardrobe worker on this job’s real tasks. So we publish no parity number. Guessing one would be worse than leaving the box empty.
What would settle it is specific. A timed test of a robot assisting a quick change on a live stage. A trial of automated garment inspection catching failures a dresser would catch. A measured comparison of costume inventory and repair tracking handled by software against a human wardrobe department over a full run. Until something like that exists and is published, the fair answer is that this job has not been measured head to head. Our full approach is on the methodology page.
Good to know: a low evidence grade is not a safety claim; it means the question has not been tested, in either direction.
When could the work change?
Most likely after 2043 (8 in 10 of our scenarios). The replacement-year method page sets out what that window does and does not measure.
Two things could pull it earlier. First, general-purpose robots with real finger dexterity: the physical share of this job’s tasks is high, and our robotics tier for it is a dexterous humanoid, which is the hardest class to build. Second, cost. The software side of the job is already cheap to automate; the hardware side is not, and the gap between the two is the whole story.
Two things hold it back. Live performance has no second take, so the tolerance for a failed quick change is near zero, and unions and production managers decide who stands in the wings. And costumes are not standard objects: hook-and-eye closures, boning, wigs and period fasteners vary by show and by body. Employment in the occupation is small, about 6,510 US jobs, with projected growth of 5.9% through 2035 and median pay near $50,400 (BLS, 2025). A small, varied, physical occupation is a weak target for expensive hardware. If robotics interests you, the guide on humanoid robots and physical jobs goes deeper.
How to stay needed in wardrobe
Lean into the parts of the job that happen in the room. Quick changes under pressure, hands-on alteration and repair, and the care and handling of fragile garments are the three tasks that keep a dresser on the call sheet. Being the person a nervous performer trusts at the five-minute call is part of that, and it is not a software feature.
Two skills compound. One is fitting and sewing at a level above basic mending, so you can solve a problem a designer did not plan for. The other is running the wardrobe paperwork with AI tools instead of against them: inventories, costume plots and repair logs drafted quickly, checked by you. That moves you toward supervisor work rather than away from it.
Nearby jobs worth comparing: Locker Room, Coatroom, and Dressing Room Attendants, Makeup Artists, Theatrical and Performance, and Tailors, Dressmakers, and Custom Sewers. You can also see how this job sits inside entertainment attendants and related workers, or across the wider arts and entertainment sector.
Next step: put this job beside another one on the compare page, or scan the list of jobs that mostly need a person to see which work holds up best.