EXPOSED
The work is overwhelmingly physical and in-person: setting up hydrotherapy and traction equipment, cleaning treatment areas, helping patients on and off tables, walking them between exercise stations, launder and stock. Almost none of that is text or screen work AI can touch, and no robot handles an unsteady post-surgical patient in a busy clinic. The exposure is on the other side: aides are unlicensed, make no clinical calls, and their clerical share — scheduling, insurance paperwork, note transcription for the PT — is exactly what AI absorbs first, which trims hours rather than eliminating the role.
Roughly flat across the period, with year-to-year wobble.
Median pay $27,000 → $35,240 +4.4% 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
+2.8% 45,600 → 46,900 on the projections basis
Growing, and only partly exposed
The BLS expects +2.8% more of these jobs by 2034, and at 48/100 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.
~6,600 openings a year on average, including replacing people who leave.
AideTherapy AideOutpatient AidePhysiotherapy AideRehabilitation AideRestorative Aide (RA)Outpatient Therapy AideRehabilitation AttendantPhysical Therapy Aide (PTA)Rehabilitation Therapy AidePhysical Therapist Aide (PTA)PT Tech (Physical Therapy Technician)PT Tech (Physical Therapist Technician)PT Attendant (Physical Therapy Attendant)Rehabilitation Technician (Rehabilitation Tech)
Holding it up: embodiment . Weakest point: liability shield .
Tasks largely resist digitisation At 16 rather than 19, because guarding a patient during parallel-bar gait training, wrapping hot packs, and transferring someone from wheelchair to mat table cannot be done remotely — but the front-desk block of your shift (booking follow-ups, chasing authorizations, typing the PT's dictated notes into the EMR) is genuinely automatable, which is what keeps this off the ceiling.
Hands-on in uncontrolled environments 18 reflects that nearly every duty happens with hands on equipment or on a person in a room that changes hour to hour — spotting an unsteady total-knee patient, dragging ultrasound units and gait belts between bays, mopping up after whirlpool — and the only reason it isn't 20 is that the clinic floor is a controlled indoor space, not a roof or a roadside.
No licence, no signature requirement 2 because no state licenses PT aides — you work under the PT's or PTA's direct supervision, your interventions are not separately billable under Medicare Part B, and if a patient falls during a transfer the claim lands on the supervising therapist and the clinic, not on a credential you hold.
Executes defined procedures on defined inputs 4 because the PT's plan of care specifies the exercise, the reps, and the modality settings; your calls are narrow ones — pausing when a patient reports pain and reporting it up the chain — and you are not permitted to modify a program, progress a patient, or interpret a response as a clinical finding.
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 (16/20 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 (2/20) is whether the law requires a licensed human to sign. Trust premium (8/20) is whether buyers specifically pay for a person. Judgment and accountability (4/20) 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 14 of this occupation's 48 points (29%).
Embodiment (18/20) 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.
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.
The clerical tier (scheduling, prior-auth forms, transcribing the PT's note) is the part AI absorbs; if clinics fully hand that to ambient documentation tools like Nuance DAX or Heidi, the remaining defined job is entirely hands-on patient transfer, gait assistance, equipment setup and turnover — a task mix with almost no automatable residue. This raises the score while reducing total hours per aide.
State practice acts that convert 'aide' from unregulated helper into a registered/credentialed support role with a defined scope and mandatory line-of-sight supervision — the model already in Texas, Louisiana and several other boards' rules on unlicensed personnel. If more boards adopt registration plus a named supervising PT whose license is on the line for delegated tasks, the role acquires a thin but real regulatory anchor that prevents substitution by untrained or unstaffed labor.
Insurer or CMS conditions of participation requiring a documented trained human present for patient transfers and ambulation in outpatient PT settings — analogous to safe-patient-handling mandates already legislated in states like California (AB 1136) and Washington for hospitals. If extended to outpatient rehab, staffing a trained aide becomes an audited requirement rather than a cost choice.
Explicit board or employer scope language making the aide responsible for recognizing and escalating red flags during exercise (syncope, chest pain, fall risk change) rather than merely executing, paired with required competency sign-off. This exists informally; codifying it in written delegation protocols makes the escalation call an owned consequential judgment.
The limit. Even with every lever, this caps in the low-to-mid 60s. Embodiment is already near maximum and cannot rise. Trust premium has no realistic route: patients pay for the licensed PT's name on the plan of care and rarely know an aide from an assistant, so there is no buyer willing to pay a premium specifically for a human aide. The core vulnerability is not machine capability but that the role is cheap, unlicensed, and hour-trimmable — automation of the clerical share cuts hours per aide even as the physical work stays untouched.
| New York-Newark-Jersey City, NY-NJ | 6,030 | $35,320 +0% |
| Los Angeles-Long Beach-Anaheim, CA | 2,720 | $37,280 +6% |
| Houston-Pasadena-The Woodlands, TX | 1,280 | $33,060 -6% |
| Phoenix-Mesa-Chandler, AZ | 1,070 | $35,010 -1% |
| Dallas-Fort Worth-Arlington, TX | 990 | $35,570 +1% |
| Philadelphia-Camden-Wilmington, PA-NJ-DE-MD | 990 | $32,240 -9% |
| Boston-Cambridge-Newton, MA-NH | 840 | $37,440 +6% |
| Detroit-Warren-Dearborn, MI | 800 | $28,570 -19% |
| Abilene, TX | 70 | $49,790 +41% |
| Seattle-Tacoma-Bellevue, WA | 400 | $46,460 +32% |
| San Francisco-Oakland-Fremont, CA | 680 | $44,440 +26% |
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 48. 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.