EXPOSED
This is hands-on work — assembling patient meal trays to dietary orders, pushing carts through hospital wings and senior-living dining rooms, delivering hotel room service, refilling steam tables — and language AI does essentially none of it. The real pressure is robotics and layout redesign: tray-delivery robots, cart automation, and centralized cook-chill kitchens already cut delivery headcount in large hospitals, and self-serve micro-markets replace attended service lines. No licensure protects the role and the judgment content is thin, so protection rests almost entirely on the physical, in-person nature of the work.
Dipped in 2020, then grew past where it started.
Median pay $24,430 → $35,360 +15.8% in real terms
This line is counted by the Bureau of Labor Statistics — the one figure on this page that isn't a judgement of ours. Headcount moves on demand, offshoring, demographics and the business cycle, and automation is one term among several, often not the loudest.
So a falling line is not evidence that AI did it, and a rising one is not evidence that it won't. Both happen in this register: some occupations resist automation and shrink anyway, others are highly automatable and keep growing. The marked year is 2020.
BLS projection, 2024–2034
+3% 277,200 → 285,300 on the projections basis
Growing, and only partly exposed
The BLS expects +3% more of these jobs by 2034, and at 47/100 the work is only partly exposed — some tasks are automatable, the core of the job is not. Nothing here is in tension.
Different clocks. The score is what current AI could do to this work today. The projection is how many of these jobs will exist in 2034. Everything between the two — how fast employers actually adopt, whether demand grows in the meantime — is why they can point opposite ways without either being wrong.
~48,000 openings a year on average, including replacing people who leave.
CurberCar HopBoat HopCurb HopFood PorterFood RunnerFood ServerLine ServerRoom ServerTray ServerCar AttendantCurb AttendantHot Dog VendorHot Tamale ManKitchen RunnerRoom AttendantHot Tamale WorkerTeletray OperatorDining Room ServerRoom Service ClerkFood Cart AttendantFood Service WorkerOutside Food ServerRoom Service Server
Holding it up: embodiment . Weakest point: liability shield .
Tasks largely resist digitisation Loading a diet-restricted tray, catching that a NPO patient's slip is still active, knocking on a room door and setting the tray within reach of someone in traction — these are sequences no software can execute, and the 14 rather than 18 reflects that tray-line assembly and cart routing are exactly the repetitive, mapped-out steps hospital cook-chill systems and delivery robots are already taking over.
Hands-on in uncontrolled environments You are on your feet the whole shift pushing 300-pound loaded carts onto elevators, working around visitor traffic and crash carts in corridors, reaching over bed rails, hauling hotel room-service trays down guest hallways, and lifting steam-table pans — uncontrolled, variable spaces, which is why this sits at 16 rather than the low teens of a fixed cafeteria counter.
No licence, no signature requirement A food handler card costs a morning and a fee in most states, and it protects nobody's job: the dietitian signs the diet order and the food-service director owns the health-code inspection, so the 3 is for that card and nothing more.
Executes defined procedures on defined inputs Your calls are bounded by the diet card, the tray ticket, and hold-time rules — flagging a wrong tray or an untouched meal to the nurse matters, but you escalate rather than decide, and no ambiguous high-stakes judgment lands on you.
The verdict above describes this occupation as a whole. Almost nobody does the typical version of a job — tick what's actually in your week and see how your own mix sits.
Your task mix speaks to task resistance (14/20 here) — how much of the day's work current AI already does. That is the dimension the boxes above are about.
It cannot move the other three. Liability shield (3/20) is whether the law requires a licensed human to sign. Trust premium (9/20) is whether buyers specifically pay for a person. Judgment and accountability (5/20) is whether the role exists to own consequential calls. Those are facts about the occupation's standing, not about which tasks are in your week — a paralegal who does only trial exhibits still holds no licence. Together they are 17 of this occupation's 47 points (36%).
Embodiment (16/20) is also a property of the work rather than the worker, but we don't tag individual tasks as physical or not, so the picker can't tell you anything about it. That's a limit of this tool, not a claim.
Did we get the list right? Tell us what's missing — the tasks are written from the outside, and you're reading this from the inside.
Cooks, Restaurant EXPOSED
Bartenders SAFE
The moves above are yours to make. This is the other half: what would have to change in the world for the occupation itself to score higher. None of it is in any one person's gift, but it is where the floor actually comes from. Scores here are not a one-way ratchet. Only two of the five dimensions — task resistance and embodiment — track what machines can do. The other three track law, what buyers will pay for, and who is answerable, and those move in both directions, often in response to the same pressure AI creates. If every lever below landed, this occupation would score around 60/100, still EXPOSED.
Task-mix shift as robots take corridor transport: the remaining job becomes bedside interaction — verifying the patient's identity band against the tray ticket, checking that the delivered tray matches the current diet order (NPO status changes after a tray is plated), and observing/recording intake. Hospital 'meal tray checkpoint' protocols already assign these to a human at the point of delivery; CMS long-term-care rules (F-tag 692) on weight loss and intake monitoring make the observation step auditable.
Formal designation of the tray-delivery worker as the last-line diet-order verifier and choking/dysphagia-risk observer. If facility policy or a Joint Commission-driven change makes the deliverer responsible for refusing a tray that conflicts with an allergy or thickened-liquid order, and for escalating a resident who cannot self-feed, the role owns a consequential call rather than executing one.
Senior-living and memory-care marketing that sells human dining companionship as a clinical anti-loneliness intervention, priced into the rate — CMS and state agencies increasingly treat social isolation as a quality measure. Buyers here are families and facility operators, not the resident, so the willingness to pay is real but narrow; it does not extend to hotel room service or hospital cafeteria lines.
Hospital and senior-living facility layouts that resist robot delivery — older multi-wing buildings with non-standard elevators, patient rooms requiring bedside tray setup and repositioning, and infection-control rules (e.g., contact-precaution/isolation rooms) that forbid shared robotic carts entering. If CMS or state health-department surveys formalize isolation-room handling requirements that only gowned staff can perform, the physical tier gets harder to strip out.
State food-handler and allergen-awareness certification is the only real hook; some states (e.g., Illinois allergen training mandates) require a certified person on site. This does not create personal liability for a server and there is no realistic route to a signature-and-liability regime for tray delivery.
The limit. Realistic ceiling is around 60. The three institutional dimensions are structurally capped: no licensure exists or is being proposed, and the trust premium is confined to the senior-living segment. Everything above the low 50s depends on the healthcare/eldercare share of the occupation, where diet verification and intake observation can be formalized into real accountability. Hotel room service, micro-market attendants, and cafeteria steam-table work have no such route and stay exposed to layout redesign, which removes the job without needing any robot to do it well.
| New York-Newark-Jersey City, NY-NJ | 19,450 | $39,120 +11% |
| Chicago-Naperville-Elgin, IL-IN | 12,200 | $35,670 +1% |
| Los Angeles-Long Beach-Anaheim, CA | 11,850 | $38,350 +8% |
| Philadelphia-Camden-Wilmington, PA-NJ-DE-MD | 7,610 | $33,300 -6% |
| Minneapolis-St. Paul-Bloomington, MN-WI | 6,740 | $36,420 +3% |
| Washington-Arlington-Alexandria, DC-VA-MD-WV | 5,760 | $37,540 +6% |
| Miami-Fort Lauderdale-West Palm Beach, FL | 4,980 | $34,950 -1% |
| Seattle-Tacoma-Bellevue, WA | 4,800 | $40,640 +15% |
| Kahului-Wailuku, HI | 200 | $62,960 +78% |
| San Jose-Sunnyvale-Santa Clara, CA | 1,960 | $46,210 +31% |
| Hanford-Corcoran, CA | 50 | $45,030 +27% |
We have no reported case of a named organisation automating this occupation. Not one deployment, not one announcement.
That is worth saying out loud next to a score of 47. The verdict above is about what the work exposes — what current AI could do to these tasks. It is not a claim that anyone has done it. For this occupation those two things have come apart completely: the capability argument is on this page, and the evidence column is empty.
Has AI actually changed your work? One tap, anonymous, and the running tally is public. Nothing else is asked of you.
Rather than check back: get the digest and we'll tell you what changed — or watch a single occupation from its own page.