← Risk register SOC 25-3041 · reviewed 2026-08-11

Tutors

175,070 US workers · median $43,350/yr · Education

EXPOSED

The informational core of tutoring — explaining a concept, working a sample problem, generating practice sets, reviewing an essay draft — is exactly what chatbots do for free, and homework-help demand is already leaking to them. What survives is the part parents actually pay for: a human who shows up weekly, notices the kid is anxious rather than confused, enforces accountability, and reports honestly to the family. Scored at the modal worker (private/agency academic tutor); credentialed special-education and dyslexia-intervention tutors sit meaningfully safer.

10-year outlook: Volume for generic homework help and content review shrinks hard by the mid-2030s, while in-person intervention, test-prep coaching, and executive-function work hold or grow at higher hourly rates.

Score — 37/100 resistance

Five dimensions, 0–20 each, summed. Higher means more protected. The arithmetic is shown so you can check it: 7 + 8 + 1 + 14 + 7 = 37.

Task resistance 7/20

Mixed — a routine tier and a judgment tier. Explaining the quadratic formula, generating twenty more practice problems, and marking up a five-paragraph essay are the bulk of a tutoring hour and an LLM does all three on demand, which caps this at 7; the residue that holds — sitting across from a 14-year-old who says 'I get it' when he doesn't, and pacing a session against a Thursday test — keeps it out of the 0-6 band.

Embodiment 8/20

Some physical or field component. Most sessions run at a kitchen table, a library branch, or a learning-center cubicle, with driving between clients, printed worksheets, and manipulatives for younger students — but nothing physical is being done to anything, and the same session already runs over Zoom without loss, which puts it at 8 rather than up with in-home caregivers.

Liability shield 1/20

No licence, no signature requirement. No state licence, no exam, no continuing-education hours: an agency can hire a strong undergrad on a background check alone and Wyzant will list anyone who passes a subject quiz, so the 1 reflects that nothing legally stops a parent from replacing you with software or a cheaper tutor tomorrow.

Trust premium 14/20

The human relationship is the product. Parents keep paying $60-90 an hour after the concept is explained because the specific person their kid will actually sit still for is the product — referrals move nearly all private business, and losing a tutor mid-semester means restarting the relationship, not just the curriculum, which earns 14; it stops short of 17+ because families do swap tutors between subjects and school years.

Judgment & accountability 7/20

Meaningful discretion. You decide whether to reteach fractions before tackling algebra, when to tell a parent their child is behind, and how to handle a student who's clearly copying answers — real calls made alone in the room — but they're bounded by the school's syllabus and the test date, and you don't diagnose, place, or write anything that follows the child on record, so 7 sits at the floor of genuine discretion.

Confidence: medium · reviewed 2026-08-11 · how scoring works

Tasks already automatable

What survives

Active moats: trust, physical-presence

How to future-proof this job

Escape hatches — adjacent fields with better verdicts

Computed from U.S. Dept. of Labor O*NET skill profiles: high overlap with what you already do, materially higher resistance score.

Special Education Teachers, Middle School SAFE · 76/100 · you already have ~79% of the skill profile

Skills to close: Coordination, Systems Evaluation, Time Management, Service Orientation

Secondary School Teachers, Except Special and Career/Technical Education SAFE · 69/100 · you already have ~79% of the skill profile

Skills to close: Quality Control Analysis, Management of Material Resources, Coordination, Management of Financial Resources

Elementary School Teachers, Except Special Education SAFE · 72/100 · you already have ~76% of the skill profile

Skills to close: Coordination, Systems Evaluation, Quality Control Analysis, Time Management

Field report — do you do this job?

Has AI actually changed your work?

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

From people who do this job

Nobody has filed one yet. If you do this work, you know things the rubric can't see.

What has actually changed in your work?

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.