Free sampler. Re-scored every six months.
Classroom teaching scores 3.6 out of 10 for AI exposure, where 10 is most at risk — one of the lower scores in this index. Special education scores 3.1. But online instruction and content-based tutoring score 6.8, nearly double. A classroom is not an information-delivery problem, which is exactly why teaching through a screen is on a completely different trajectory.
The short answer for parents: yes, and AI is on balance good news for classroom teachers. It automates the marking, planning and paperwork that drive people out of the profession, while leaving the part that drew them in. The caution is that education-adjacent careers built on content delivery share the online track's exposure, not the classroom's protection.
Every career in this index is scored 1–10, where 10 is most exposed to AI. Same six factors, same weights, applied identically to a teacher and a paramedic.
| Teaching track | 2023 | 2025 | Now | 3-yr move | Band |
|---|---|---|---|---|---|
| Special education | 2.7 | 2.9 | 3.1 | +0.4 | Low |
| Elementary / early years | 3.2 | 3.4 | 3.6 | +0.4 | Low |
| Secondary — STEM subjects | 3.4 | 3.6 | 3.8 | +0.4 | Low |
| Secondary — general subjects | 3.7 | 3.9 | 4.1 | +0.4 | Low–Mod |
| Online instruction / tutoring | 5.5 | 6.2 | 6.8 | +1.3 | Mod–High |
2023 and 2025 figures are reconstructed using current methodology, not archived from past editions.
For scalethe median career in this edition scores around 5.5. Bedside nursing scores 2.8. Entry-level software development scores 8.1.
Classroom teaching's +0.4 over three years is tied for the slowest movement in the index. Online instruction moved +1.3 over the same period — the same skill, delivered through a screen, on a completely different trajectory. No pair we score shows more clearly that setting is destiny.
Not in a classroom. AI tutors are already good at explaining content patiently and infinitely — which turns out to be a small fraction of what a teacher does.
Adoption is already near-universal — the large majority of teachers and students used AI in the last school year, and most teachers who did reported it improved their teaching and gave them more time with students.
For a profession with a retention crisis, that matters. The tasks being automated are the ones people cite when they leave.
How elementary classroom teaching rates against each. Ratings are 0–10 on each factor's own terms.
So you can see what the analysis actually looks like.
How much of this job can AI already do? — rated 7.7
This is the factor that carries the most weight in our model at 35%, and teaching is the clearest case in the whole index of it being outweighed.
Teaching rates 7.7 on automatability — higher than nursing, higher than medicine, higher than most careers we score. A very large share of a teacher's working hours goes to planning, resource creation, differentiation, marking, feedback and administrative reporting. Nearly all of it is now automatable, and much of it is already being automated in practice.
If our model let automatability alone drive the score, teaching would sit in the moderate-to-high band alongside marketing and business analysis.
It doesn't, because the other five factors run strongly the other way. Presence rates 9.0 — a classroom is embodied, social work. Trust rates 9.5, because parents hand over their children, which is not a relationship anyone delegates to software. Certification is mandated, and the entry route requires classroom practice. And schools perform a custodial and social function that has nothing to do with information transfer at all.
The design choice that produces this result is deliberate. Our three protection factors carry 50% of the weight in total — more than automatability's 35%. That means protection can outweigh exposure. If we had weighted exposure more heavily, teaching would score high and the model would be wrong.
And it is exactly why the online track scores 6.8. Take the identical subject knowledge and identical automatable tasks, strip out presence, weaken trust, remove certification requirements, and the same 7.7 automatability rating now dominates. Nothing about the work changed. The setting did.
Ranked by exposure, safest first:
The AI question here is genuinely positive, and the difficult questions are the perennial ones.
Teaching has some of the strongest structural protection in this index — certification, presence, trust, and shortages in most regions, especially in math, science, special education and bilingual instruction. Landing a first job is among the easiest of any career we score.
What has not changed: workload, where the visible hours are a fraction of the real ones. Behavior management, which is harder than most people expect. Pay that lags comparable graduate careers. And high early-career attrition.
AI plausibly improves the first of those. It does nothing for the rest.
One thing worth asking any teacher-training program: how do they prepare candidates to use these tools well? Under half of teachers report receiving any institutional training or guidance on AI, which means a newly-qualified teacher who is genuinely fluent walks into a shortage market with a real edge.
This sampler tells you where teaching stands. The full profile tells you what to do about it.
Most families are weighing two or three careers seriously, and a few more they haven’t ruled out. Pick the ones you need.
Each includes a short version written directly to the student and the technical scoring appendix. Spring 2027 updates of whatever you buy are included.
Read one complete profile free → We publish computer science in full so you can judge the depth before buying anything.