SAFE
The core of this job is being physically in the room with acutely unwell people: observing behavior, de-escalating agitation, assisting with bathing and meals, taking vitals, participating in physical restraint or seclusion, and leading group activities. None of that digitizes, and hospital robotics are nowhere close. AI will absorb the paperwork layer — shift notes, behavior logs, incident report drafting, scheduling — and predictive monitoring tools may reshape how observation rounds are documented, but they increase the need for a human body on the unit rather than reduce it.
Headcount grew steadily across the period.
Median pay $33,780 → $45,130 +6.9% 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
+20% 144,500 → 173,300 on the projections basis
Hard to automate, and growing
The work resists current AI and the BLS projects +20% 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.
~15,900 openings a year on average, including replacing people who leave.
CounselorMilieu TherapistPsychology AssociatePsychiatric AssistantResidential Aide (RA)Mental Health AssociatePatient Care SpecialistMental Health SpecialistSupport Team Member (STM)MHW (Mental Health Worker)Behavioral Health AssociatePsychiatric Technician (PT)MHA (Mental Health Assistant)Mental Health Technician (MHT)BHT (Behavioral Health Technician)Behavior Technician (Behavior Tech)RBT (Registered Behavior Technician)LPT (Licensed Psychiatric Technician)Rehabilitation Technician (Rehab Tech)Behavioral Technician (Behavioral Tech)Psychology Technician (Psychology Tech)Health Care Technician (Health Care Tech)Autism Behavior Technician (Autism Behavior Tech)Psychiatric Safety Tech (Psychiatric Safety Technician)
Holding it up: embodiment . Weakest point: liability shield .
Tasks largely resist digitisation Fifteen-minute safety checks, one-to-one observation of a patient on suicide watch, talking someone down from a doorway blockade, and a four-point takedown all require a person who can physically intervene within seconds — the only genuinely automatable share is the charting and med-count paperwork, which is why this sits at 16 rather than a perfect 20.
Hands-on in uncontrolled environments You work an unlocked-nothing unit where patients spit, swing, and elope, you help shower someone who is actively psychotic, and you put hands on a body during restraint — an uncontrolled environment with an unpredictable human at the center, which is as embodied as healthcare gets short of field EMS.
Certification preferred, not legally required Most states license or certify psych techs at some level and a few (California's Psychiatric Technician license) make it a real credential with board discipline, but the treatment plan, the restraint order, and the medication decision are all signed by an RN or psychiatrist — you document and execute rather than authorize, which caps this at 7.
Meaningful discretion You decide in the moment whether pacing is a warning sign, whether to call a code or keep talking, and when a seclusion is warranted — real calls with real consequences, but made inside standing orders, restraint protocols, and a nurse you can summon down the hall, which is why this lands at 11 rather than in the high teens.
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 (7/20) is whether the law requires a licensed human to sign. Trust premium (14/20) is whether buyers specifically pay for a person. Judgment and accountability (11/20) 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 32 of this occupation's 67 points (48%).
Embodiment (19/20) 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 81/100, still SAFE.
Task-mix shift: if AI absorbs behavior logging, incident report drafting and rounds documentation, the residual role concentrates on the ambiguous calls — reading escalation before it happens, deciding when to touch a patient, when to call a code. This tier already exists in the job and is the part that survives.
Same task-mix mechanism: the paperwork tier is the automatable tier and it is a real share of the shift. Its removal raises the resistance of what remains without any new law.
State licensure/certification for psychiatric technicians expanding beyond the four states that currently license (California, Colorado, Kansas, Arkansas) — with a scope-of-practice rule making the certified tech the named accountable observer on seclusion/restraint documentation, personally subject to board discipline. CMS Conditions of Participation §482.13 already requires trained staff to conduct and document 1:1 monitoring during restraint; naming a licensed role rather than 'trained staff' would harden this.
A Joint Commission or state licensing rule requiring that AI-generated behavioral observation entries and predictive-agitation alerts be countersigned by a named human staff member who attests to direct visual observation — creating a chain-of-custody signature that cannot be automated.
Mandated minimum-staffing and assignment-accountability laws for behavioral health units (the pattern in California AB 394 and Oregon HB 2697, which name specific units and ratios) that assign the tech, not just the RN, formal responsibility for the restraint/de-escalation decision path and post-incident debrief.
The limit. Trust premium has no realistic route up. Buyers here are hospitals, state facilities and correctional systems purchasing on ratio compliance and cost, not patients choosing a technician; nobody pays extra for a specific human in this role and no credentialing body is positioned to create that market. Embodiment is already near ceiling at 19 and cannot meaningfully rise. Note also that a rising score here does not mean rising wages — the same institutional pressure that hardens liability tends to arrive alongside understaffing.
| Chicago-Naperville-Elgin, IL-IN | 6,730 | $46,060 +2% |
| Phoenix-Mesa-Chandler, AZ | 6,630 | $46,550 +3% |
| Los Angeles-Long Beach-Anaheim, CA | 4,770 | $48,040 +6% |
| Philadelphia-Camden-Wilmington, PA-NJ-DE-MD | 4,560 | $46,920 +4% |
| Atlanta-Sandy Springs-Roswell, GA | 4,230 | $47,650 +6% |
| Miami-Fort Lauderdale-West Palm Beach, FL | 3,890 | $44,500 -1% |
| Boston-Cambridge-Newton, MA-NH | 3,460 | $57,610 +28% |
| Houston-Pasadena-The Woodlands, TX | 3,300 | $43,320 -4% |
| Stockton-Lodi, CA | 330 | $95,580 +112% |
| Napa, CA | 320 | $83,960 +86% |
| Visalia, CA | 200 | $76,700 +70% |
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 67. 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.