SAFE
The paperwork half of this job — progress notes, treatment plan drafts, intake summaries, insurance authorization narratives, screening instrument scoring — is already being absorbed by AI scribes and documentation tools. The other half, sitting with a client in withdrawal or crisis, holding a relapse confrontation, judging when to escalate to involuntary commitment or call a probation officer, is exactly what buyers and courts require a credentialed human for. Most states require LPC/LCSW/LADC licensure with personal accountability for clinical decisions, and much of the work happens in-person in clinics, residential programs, jails, and hospitals.
Headcount grew steadily across the period.
Median pay $46,240 → $59,350 +2.7% 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
+16.8% 483,500 → 564,600 on the projections basis
Hard to automate, and growing
The work resists current AI and the BLS projects +16.8% 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.
~48,300 openings a year on average, including replacing people who leave.
ClinicianCounselorCase ManagerDrug CounselorElder CounselorGrief CounselorGroup CounselorBehavior AnalystBehavior TherapistBehavioral AnalystClinical CounselorLicensed ClinicianLicensed CounselorLicensed TherapistQuitline CounselorAddiction CounselorAddiction TherapistAlcoholic CounselorTreatment CounselorBehavioral ClinicianBehavioral TherapistDrug Abuse CounselorGroup Home CounselorOutpatient Counselor
Holding it up: trust premium . Weakest point: embodiment .
Mixed — a routine tier and a judgment tier ASAM criteria placement interviews, motivational interviewing with a client who is minimizing use, group process facilitation, and relapse-confrontation sessions don't reduce to text generation — but DSM-5 criteria checklists, ASI and AUDIT-C scoring, treatment plan goal libraries, and the weekly progress note cycle do, which is why this sits at 13 rather than 17.
Some physical or field component You are in the room for withdrawal monitoring, urine collection observation, group rooms in residential programs, jail visitation booths, and hospital bedside consults, and sometimes physically intervening in a de-escalation — but you do not perform procedures or lift patients, so the physical demand is presence and proximity rather than a 15-plus trade skill.
Licensed human required and personally liable LPC, LCSW, LMHC, or LADC/CADC licensure is a hiring condition in nearly every state, board complaints for boundary violations or failure-to-warn go against your name and not the agency's, and 42 CFR Part 2 confidentiality breaches expose you personally — the 14 rather than 18 reflects that a meaningful share of counselor-level SUD work is still done by certified-but-unlicensed staff under a clinical supervisor's signature.
Exists to be accountable for ambiguous calls You decide when a disclosure triggers Tarasoff duty-to-warn, when to file for emergency commitment, whether a positive screen goes in the report that sends someone back to a probation revocation hearing, and when someone in acute withdrawal needs medical transfer — high-stakes calls made on ambiguous evidence in real time, though most sit inside ASAM and agency protocols rather than being wholly unstructured.
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 (13/20 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 (14/20) is whether the law requires a licensed human to sign. Trust premium (18/20) 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 46 of this occupation's 69 points (67%).
Embodiment (10/20) 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 82/100, still SAFE.
State licensing boards or federal rules explicitly barring AI from performing clinical assessment/treatment planning without a licensed clinician's independent review and signature — already in motion: Illinois' Wellness and Oversight for Psychological Resources Act (HB 1806, signed Aug 2025) bans AI-delivered therapy and requires licensed-professional oversight; Nevada AB 406 and Utah HB 452 similar. If more states adopt and boards begin disciplining unsigned AI-generated treatment plans, personal liability attaches more firmly.
Task-mix shift: as scribes absorb notes, authorizations, and instrument scoring, the residual role concentrates in duty-to-warn (Tarasoff) calls, involuntary-commitment petitions, court and probation reporting, and child-welfare mandated reporting — all decisions with named-clinician consequences. Recognizable if caseload standards or billing codes shift toward crisis/assessment units and away from documentation time.
Peer-support and 12-step-adjacent norms already price human sobriety experience explicitly; formal certification of peer specialists and lived-experience requirements in state Medicaid billing make the human identity of the provider a reimbursable attribute rather than incidental.
42 CFR Part 2 / DEA telehealth rules for buprenorphine and OTP dosing tied to a named licensed counselor of record for the required counseling component, plus CARF/Joint Commission accreditation standards specifying that AI documentation cannot substitute for the credentialed clinician's attestation.
Drug court and problem-solving court rules naming a licensed counselor as the clinical recommender on sanction/termination decisions, where the judge relies on that clinician's testimony.
Payer and accreditation rules requiring that AI-drafted notes be independently verified against a live session, plus restrictions on chatbot-delivered intervention (FTC/state AG actions against AI companion and therapy apps), pushing the automatable tier back into supervised rather than replaced work.
The limit. trust_premium at 18 and embodiment are near their practical ceilings — embodiment cannot rise much since the physical component is presence, not manipulation, and presence is what the trust premium already prices. Realistic upside is roughly 75-80, not higher. Counterpressure: Medicaid rate compression and workforce shortages create political appetite for AI-extended caseloads, which cuts the other way on task_resistance.
| Los Angeles-Long Beach-Anaheim, CA | 26,960 | $51,910 -13% |
| New York-Newark-Jersey City, NY-NJ | 26,200 | $61,020 +3% |
| Chicago-Naperville-Elgin, IL-IN | 15,580 | $61,600 +4% |
| Philadelphia-Camden-Wilmington, PA-NJ-DE-MD | 14,920 | $58,140 -2% |
| Boston-Cambridge-Newton, MA-NH | 11,150 | $62,130 +5% |
| Miami-Fort Lauderdale-West Palm Beach, FL | 10,240 | $57,290 -3% |
| San Francisco-Oakland-Fremont, CA | 9,520 | $68,400 +15% |
| Washington-Arlington-Alexandria, DC-VA-MD-WV | 9,030 | $65,270 +10% |
| Fairbanks-College, AK | 100 | $83,680 +41% |
| Anchorage, AK | 810 | $81,120 +37% |
| San Luis Obispo-Paso Robles, CA | 350 | $80,770 +36% |
KQED reports San Francisco officials criticising Kaiser Permanente's use of AI in mental health care delivery.
KQED reports Northern California Kaiser Permanente mental health therapists held a one-day strike over concerns about AI use and patient care.
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.