SAFE
This residual category holds art, music, dance/movement, recreational and similar therapists who work body-to-body and instrument-in-hand with patients — stroke survivors relearning gait to rhythm, dementia patients reached through song, trauma clients who can't yet talk about it. The automatable slice is the paperwork: progress notes, treatment plan drafting, insurance documentation, outcome-scale scoring. The core session — reading a nonverbal patient's affect in real time and adjusting the intervention on the spot — has no screen equivalent, and most states or national boards require credentialed practitioners for reimbursement.
Dipped in 2020, then grew past where it started.
Median pay $52,650 → $77,930 +18.4% 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
+11.5%
Percentage only. The projection counts a different population from the 22,640 above — it includes self-employed workers, which for this occupation is most of them, so the two headcounts are not comparable.
Hard to automate, and growing
The work resists current AI and the BLS projects +11.5% more of these jobs by 2034. Note that safe does not mean well paid — several of the fastest-growing resistant occupations are among the lowest paid on the register.
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.
~4,100 openings a year on average, including replacing people who leave.
DoctorRolferHerbalistOsteopathTherapistSexologistHerb DoctorEcotherapistNephrologistArt TherapistKinesiologistSex TherapistAromatherapistEar SpecialistHeliotherapistHerb CounselorHydrotherapistHypnotherapistPeripatologistPlay TherapistClinical FellowColon TherapistGroup TherapistMagnetic Doctor
The BLS uses Therapists, All Other for work that doesn't fit any named occupation, so it covers roles that have little in common with each other. Two consequences worth knowing before you read anything below:
Holding it up: trust premium . Weakest point: liability shield .
Tasks largely resist digitisation A drum circle with six psychiatric inpatients, a hemiparetic patient's hand physically guided through a clay-forming task, or a recreational therapist's community-reentry outing to a grocery store are improvised minute-by-minute from what the patient does next — only the notes, MDS/FIM scoring, and treatment-plan boilerplate sit in the automatable 5 points.
Hands-on in uncontrolled environments You are in the room, often supporting weight, positioning instruments, spotting a patient on an adaptive bike, or being physically adjacent to someone who may become agitated — but it's a treatment room, day room, or facility grounds rather than a roadside or a rooftop, which is what keeps this at 15 rather than 19.
Licensed human required and personally liable Board credentials (ATR-BC, MT-BC, CTRS, BC-DMT) plus state licensure in a growing number of jurisdictions gate reimbursement and your own name is on the note, but many of these titles are certifications rather than a practice act, so no statute makes it illegal for an uncredentialed person to run the activity — that gap holds this at 11 instead of 16.
Exists to be accountable for ambiguous calls You decide in the moment whether a patient's flat affect means disengagement or an emerging medical change, whether to stop a session escalating toward self-harm, and whether a resident is safe for an unsupervised community outing — calls you make alone and defend to the interdisciplinary team afterward.
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 (15/20 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 (11/20) is whether the law requires a licensed human to sign. Trust premium (17/20) is whether buyers specifically pay for a person. Judgment and accountability (14/20) 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 42 of this occupation's 72 points (58%).
Embodiment (15/20) 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 84/100, still SAFE.
State-level licensure (not just national board certification) for creative arts therapists spreading beyond the ~10 states that have it — e.g. more states adopting NY-style licensed creative arts therapist (LCAT) titles, or music therapy licensure bills of the kind AMTA/CBMT's State Recognition Operational Plan has been filing yearly. Licensure makes a named practitioner the legally accountable signer on treatment plans and makes Medicaid/private reimbursement contingent on that signature.
Task-mix shift: this job genuinely has two tiers. If AI documentation tools absorb progress notes, treatment-plan drafting, insurance authorization letters and outcome-scale scoring, the residual day is almost entirely in-room clinical improvisation, which raises the share of the day AI cannot do. Watch for EHR vendors (Epic, WebPT) shipping arts/rec-therapy note templates with ambient scribing.
Formal inclusion of arts/rec therapists on hospital interdisciplinary care teams with charting authority over risk decisions — e.g. documenting suicidality disclosures surfaced in a nonverbal session, or clearing a stroke patient for weight-bearing gait-to-rhythm work. Joint Commission behavioral health accreditation standards that name these disciplines in the treatment team would formalize the call.
CMS or Medicare Advantage plans adding recreational/creative arts therapy as a separately billable service with a supervising-credentialed-provider attestation requirement (analogous to the qualified therapeutic recreation specialist requirement already in the SNF Conditions of Participation, 42 CFR 483.24 activities provision) extended to more settings.
Already near ceiling at 17 — buyers of dance/movement and trauma-focused arts therapy are explicitly purchasing embodied human presence, and the modality's theory of change is the relationship itself. The realistic lever is not raising the premium but preserving the payer channel: if insurers keep covering these as skilled therapy rather than pushing them to unreimbursed 'wellness activity', the premium holds.
The limit. Embodiment and trust are effectively maxed for the in-session core; nearly all remaining headroom is in liability_shield, which is unusually low (11) only because most of these disciplines rely on national board certification rather than state licensure. Full 50-state licensure plus explicit CMS billing status is the realistic upper bound, and even then the occupation's exposure is less to AI than to being cut as a non-mandated service line in hospital budget rounds.
| Philadelphia-Camden-Wilmington, PA-NJ-DE-MD | 1,330 | $81,170 +4% |
| Atlanta-Sandy Springs-Roswell, GA | 740 | $62,450 -20% |
| Baltimore-Columbia-Towson, MD | 740 | $46,540 -40% |
| Washington-Arlington-Alexandria, DC-VA-MD-WV | 480 | $76,630 -2% |
| Dallas-Fort Worth-Arlington, TX | 370 | $75,120 -4% |
| Los Angeles-Long Beach-Anaheim, CA | 370 | $77,610 +0% |
| Minneapolis-St. Paul-Bloomington, MN-WI | 370 | $75,350 -3% |
| Houston-Pasadena-The Woodlands, TX | 290 | $63,690 -18% |
| San Francisco-Oakland-Fremont, CA | 100 | $122,120 +57% |
| Lexington-Fayette, KY | 80 | $112,070 +44% |
| Atlantic City-Hammonton, NJ | 100 | $108,650 +39% |
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 72. 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.