← Risk register SOC 21-1093 · reviewed 2026-08-11

Social and Human Service Assistants

437,860 US workers · median $45,930/yr · Social Service

EXPOSED

Roughly half the job is documentation and navigation work AI already does well: intake forms, benefits eligibility screening, case notes, referral lists, service directories, progress reports. The other half is showing up — home visits, transporting clients to appointments, sitting with someone in crisis, noticing the unsafe apartment or the missed meds — which no current system does. The occupation is unlicensed and supervised by social workers, so it has no legal shield; caseload sizes may grow as paperwork thins rather than headcount collapsing.

10-year outlook: Employment holds up because the visits and crisis work still require a body in the room, but the paperwork half of the job compresses hard — expect larger caseloads, more field time, and hiring that favors people with crisis and population-specific credentials.

US employment, 2019–2025+8.3%
404,450437,860 workers

Headcount grew steadily across the period.

Median pay $35,060 → $45,930 +4.8% in real terms (nominal +31.0%, less ~25% US inflation over the period)

The job count is not the verdict

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

+6.4% 449,600 → 478,500 on the projections basis

Growing, and only partly exposed

The BLS expects +6.4% more of these jobs by 2034, and at 50/100 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.

~50,600 openings a year on average, including replacing people who leave.

One email if this score changes. Watch as many occupations as you like from the same address — no account, and nothing is sent on a schedule, only when a verdict actually moves.

Also known as — 24 job titles this covers

Titles reported by people doing this work, from the US Department of Labor's O*NET survey. If your job title is here, this page is about your work even though the name doesn't match.

AdvocateCase AideHome VisitorService AideWelfare AideHouse VisitorIntake WorkerCase Work AideChild AdvocateCommunity AideCounselor AideCourt AdvocateSupport WorkerYouth AdvocateChildren's AideClient AdvocateCounseling AideFamily AdvocateManagement AideVictim AdvocateWelfare VisitorGerontology AideShelter AdvocateHousing Assistant

Score — 50/100 resistance

Holding it up: trust premium (14/20). Weakest point: liability shield (4/20).

Five dimensions, 0–20 each, summed. Higher means more protected. The arithmetic is shown so you can check it: 10 + 13 + 4 + 14 + 9 = 50. · Scored 2026-08-11, and re-examined when evidence accumulates rather than on a schedule.

Task resistance 10/20

Mixed — a routine tier and a judgment tier Eligibility screening for SNAP, TANF, Medicaid and housing vouchers is rule-lookup work that software already performs, and case notes and referral packets are drafting tasks — but the home visit where you see the empty fridge, the ride to the methadone clinic, and de-escalating a client who won't come to the door are not digitisable, which is why this lands mid-band rather than in the automatable tier.

Embodiment 13/20

Hands-on in uncontrolled environments You are in clients' apartments, shelters, group homes and jails on their schedule, driving them to appointments in your own car and occasionally physically assisting someone with mobility or hygiene — uncontrolled settings you don't govern — though it stops short of the 17+ range because there's no equipment operation or clinical hands-on care.

Liability shield 4/20

No licence, no signature requirement No state licence exists for this title; the supervising LCSW or LMSW signs the assessments and treatment plans and carries the board exposure, so the only thing standing between you and substitution is an employer's preference for a bachelor's degree or an HHS-funded certificate.

Trust premium 14/20

The human relationship is the product A client who has been failed by systems repeatedly will disclose the eviction notice, the relapse, or the abusive partner to you and not to the agency — that named-person continuity is why the 14 sits above the relationship-helps band, but the relationship is attached to a caseload the agency reassigns, not owned by you like a therapist's panel.

Judgment & accountability 9/20

Meaningful discretion You decide when a situation warrants a mandated child-abuse report, when to escalate to your supervisor, and how to triage a caseload of forty — real calls, but the treatment plan, the diagnosis and the case closure decision belong to the credentialed supervisor above you, which keeps this from the high-stakes-ownership band.

Confidence: high · reviewed 2026-08-11 · how scoring works · 1 deployment report on file

What this job involves — and which parts are yours

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.

AI already does these at usable quality

These still need a person

Active moats on the surviving side: embodiment, trust, physical-presence

How to future-proof this job

Training paths for your skill gaps: Learning How to Learn — the most-taken course on Coursera, and free free to audit · Khan Academy — physics, chemistry and biology from the ground up free · Coursera — people management and team leadership specialisations free to audit · MIT OpenCourseWare — problem-solving and analytical method courses free · Coursera — active listening and communication skills free to audit · Coursera — decision making under uncertainty free to audit · MIT OpenCourseWare — operations management free

All 35 skills ranked by how many jobs they open →

Where this experience transfers — occupations you could move toward

Computed from U.S. Dept. of Labor O*NET skill and knowledge profiles: high overlap with what you already do, a materially higher resistance score, no large jump in required training, and no licence you would have to start a new pipeline to get. Targets that pay meaningfully less, that are themselves COOKED, or whose own headcount is falling are excluded — a move into a shrinking trade is not an escape.

Mental Health and Substance Abuse Social Workers SAFE · 70/100 · you already have ~87% of the skill profile

Skills to close: Learning Strategies, Science, Management of Personnel Resources, Complex Problem Solving

Marriage and Family Therapists SAFE · 74/100 · you already have ~87% of the skill profile

Skills to close: Active Listening, Complex Problem Solving, Judgment and Decision Making, Management of Personnel Resources

Healthcare Social Workers SAFE · 71/100 · you already have ~85% of the skill profile

Skills to close: Operations Analysis, Management of Personnel Resources, Science, Complex Problem Solving

What would move this back up — beyond any one person

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 67/100 — SAFE.

5 specific changes that would raise this score
  • already happening task resistance +4

    If benefits screening, intake, referral matching and case-note drafting are fully automated by state agency AI (already piloted in e.g. CalSAWS-type eligibility systems and Medicaid navigation tools), the residual job becomes home visits, crisis de-escalation, and safety observation — the tier no system does. Task-mix shift, no law needed, but headcount may fall even as per-worker resistance rises.

  • already happening trust premium +3

    If peer-delivered services keep expanding in behavioral health and reentry programs, where the purchased good is explicitly lived experience and a human relationship (SAMHSA peer support funding, recovery-community organizations), the human is the deliverable rather than a channel.

  • plausible liability shield +4

    State credentialing of paraprofessional roles with personal accountability — e.g. certified peer support specialist and community health worker certification now Medicaid-billable in 30+ states; if billing rules require a named certified individual to attest to each encounter note (not just a supervising LCSW), a thin shield attaches where none exists.

  • plausible embodiment +3

    If Medicaid HCBS waiver rules and state APS protocols continue to require in-person face-to-face contact at set intervals (post-COVID telehealth flexibilities expiring for home-visit categories), the physical share of the day grows as desk work thins.

  • plausible judgment accountability +3

    If mandated-reporter statutes and child/adult-safety protocols are read to place the reporting duty on the worker who made the observation — with AI outputs treated as inadmissible substitutes for a personal safety judgment — the role owns a consequential ambiguous call rather than relaying one upward.

The limit. Ceiling is low: the role's economics are set by public agency budgets that treat it as the cheap tier under licensed social workers, so productivity gains from paperwork automation are most likely absorbed as bigger caseloads. Even with certification, the sign-off liability stays with the supervising LCSW/LCPC.

These are conditions, not forecasts — what would have to happen, not what will. Specific rules, cases and bills are named so you can go and check whether they exist and where they stand; verify before relying on any of them. Nothing here is legal or financial advice.

Where this work is, and what it pays there

BLS metro figures for 388 areas. The verdict above does not change by city — the rubric judges what the work involves, not where it happens — but pay and headcount do, and the national median hides a very wide range.

Most of these jobs

New York-Newark-Jersey City, NY-NJ 35,540 $47,860 +4%
Los Angeles-Long Beach-Anaheim, CA 18,790 $50,330 +10%
Chicago-Naperville-Elgin, IL-IN 16,080 $47,180 +3%
Philadelphia-Camden-Wilmington, PA-NJ-DE-MD 12,160 $45,370 -1%
San Francisco-Oakland-Fremont, CA 7,970 $57,580 +25%
Washington-Arlington-Alexandria, DC-VA-MD-WV 7,790 $48,370 +5%
Minneapolis-St. Paul-Bloomington, MN-WI 7,160 $49,750 +8%
San Diego-Chula Vista-Carlsbad, CA 6,400 $49,500 +8%

Best paid

Eau Claire, WI 390 $63,870 +39%
San Jose-Sunnyvale-Santa Clara, CA 1,930 $60,650 +32%
Fond du Lac, WI 180 $58,940 +28%

Percentages are against this occupation's national median of $45,930. Counts are jobs in that metro, not vacancies. Metros where the BLS suppressed the cell are absent rather than shown as zero.

Who is actually doing this

The score above is about what the work exposes. This is reporting about real deployments in this occupation — the difference between "could be automated" and "somebody automated it."

Code for America

1 of 1 reported case, with sources

Quick take — do you do this job?

Has AI actually changed your work? One tap, anonymous, and the running tally is public. Nothing else is asked of you.

Self-reported and unverified — a sentiment signal, not a survey. One response per person per occupation; you can change your answer.

Field reports — what people say has changed

No field reports yet. A written account takes a paragraph rather than a tap, goes to an editor before it appears, and is the one thing on this page the rubric cannot produce on its own.

File a field report

Concrete beats general: a tool that arrived, a task that moved, a headcount decision you watched happen. Don't include anything that identifies you or your employer if that would put you at risk.

Watch this verdict
Kept current

Rather than check back: get the digest and we'll tell you what changed — or watch a single occupation from its own page.