SAFE
The paperwork layer of teaching — lesson plans, worksheets, rubrics, first-pass essay feedback, quiz generation, parent email drafts — is already being handled well by AI, and that is real time savings, not a real threat. What is not automatable is standing in a room with 28 sixteen-year-olds, holding attention, noticing who is quietly failing, and being the adult legally responsible for them; state licensure and in loco parentis duty of care make a credentialed human non-optional. The main risk is budget-driven class-size increases and a shift of some content delivery to software, which compresses headcount rather than eliminating the job.
Roughly flat across the period, with year-to-year wobble.
Median pay $61,660 → $72,040 -6.5% 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
-1.6% 1,094,500 → 1,076,700 on the projections basis
Hard to automate, but shrinking anyway
The work resists current AI, yet the BLS projects -1.6% by 2034. Whatever is shrinking this occupation, the evidence does not point to automation — demand, demographics, offshoring and industry decline all shrink jobs that no machine could do.
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.
~66,200 openings a year on average, including replacing people who leave.
CoachTeacherEducatorArt TeacherArt EducatorBand TeacherChoir TeacherLatin TeacherMusic TeacherOrgan TeacherPiano TeacherVocal TeacherVoice TeacherArt InstructorChoral TeacherFrench TeacherGuitar TeacherHealth TeacherHebrew TeacherMusic EducatorSewing TeacherSpeech TeacherSports TeacherTyping Teacher
Holding it up: embodiment . Weakest point: task resistance .
Mixed — a routine tier and a judgment tier A 12 rather than a 16 reflects that a large slice of the week — writing unit plans to state standards, generating differentiated worksheets, grading multiple-choice and even first-draft essays, logging grades, drafting parent emails — is genuinely machine-assistable now, while the irreducible core (running a Socratic seminar, redirecting a defiant student mid-lesson, re-teaching a concept three ways because you read the room) has no digital substitute.
Hands-on in uncontrolled environments A high school classroom is an uncontrolled environment by any honest definition — 28 adolescents, hallway and cafeteria duty, breaking up a fight, chemistry lab supervision, fire drills and lockdown drills, bus and field-trip chaperoning — and the 15 rather than 19 only reflects that the intellectual work itself is done seated with a screen and a whiteboard.
Licensed human required and personally liable State teaching licensure with content-area endorsement is a hard legal gate: an unlicensed adult cannot be the teacher of record, cannot sign off on credits toward a diploma, and cannot discharge in loco parentis duty of care — the 14 rather than 18 is because districts routinely fill seats with emergency and provisional certificates when they cannot hire, and personal malpractice exposure is far thinner than a physician's.
Meaningful discretion State standards, pacing guides, district curricula and standardized testing calendars constrain a lot of the day, which caps this at 13 — but the calls that matter are yours alone and are made without a script: whether a bruise or a change in a kid's affect triggers a mandatory-reporter call, whether a plagiarism case goes to the office, whether to fail a senior short of credits, whether to escalate a suicide-risk disclosure.
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 (12/20 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 (15/20) is whether buyers specifically pay for a person. Judgment and accountability (13/20) 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 69 points (61%).
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 79/100, still SAFE.
State-level mandates that AI instructional tools cannot be the system of record for grades or IEP-adjacent accommodations without a licensed teacher's sign-off — the pattern in emerging state AI-in-education guidance (e.g. California AB 2876-style requirements, and district AI policies requiring a credentialed 'teacher of record' for any course with AI-delivered content, as in Ohio and Texas virtual-school teacher-of-record rules). Also class-size and supervision-ratio statutes written to count only licensed adults, not software-supervised students.
Genuine two-tier job: if worksheet generation, first-pass essay marking, and lesson planning are fully absorbed, the residual role is behavior management, differentiation for the 5-8 students who are failing, and family conferencing — none of which current systems do. Task resistance rises mechanically as the automatable tier is stripped out, though this compresses headcount at the same time.
Parent and union resistance to AI-delivered instruction becoming contractual: NEA/AFT locals bargaining language that AI cannot replace direct instruction hours or determine grades (AFT's AI guidance and several 2024-25 district contracts contain versions of this). Separately, private and parochial school marketing that explicitly promises small human-taught classes sustains a paid premium in the ~10% of enrollment outside public systems.
Expansion of mandated-reporter and student-threat-assessment duties: post-Uvalde state laws and district behavioral threat assessment teams (Texas SB 11, Virginia's threat assessment statute) place named teachers on teams making documented consequential calls about student risk. If AI screening tools flag students and a licensed teacher must adjudicate and document the disposition, the judgment tier is formalized rather than advisory.
The limit. Levers protect the credential and the room, not the headcount. Every mechanism above can hold while districts raise class sizes from 28 to 34 and cut positions — the job stays human and there are fewer of them. Nothing here is a lever against budget-driven compression.
| New York-Newark-Jersey City, NY-NJ | 75,780 | $100,800 +40% |
| Chicago-Naperville-Elgin, IL-IN | 32,790 | $98,970 +37% |
| Los Angeles-Long Beach-Anaheim, CA | 32,250 | $100,960 +40% |
| Dallas-Fort Worth-Arlington, TX | 28,000 | $65,890 -9% |
| Houston-Pasadena-The Woodlands, TX | 26,260 | $65,540 -9% |
| Philadelphia-Camden-Wilmington, PA-NJ-DE-MD | 25,460 | $80,560 +12% |
| Boston-Cambridge-Newton, MA-NH | 20,600 | $98,850 +37% |
| Washington-Arlington-Alexandria, DC-VA-MD-WV | 20,400 | $80,300 +11% |
| Mount Vernon-Anacortes, WA | 240 | $121,190 +68% |
| Oxnard-Thousand Oaks-Ventura, CA | 2,210 | $108,850 +51% |
| San Jose-Sunnyvale-Santa Clara, CA | 4,130 | $108,230 +50% |
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 69. 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.