School teacher
Gets thirty specific children to learn something on a specific day — and holds the room together while doing it.
This is not a probability of losing your job. It combines how much of the role's task load is exposed to automation with how far adoption has actually gone — useful for comparing occupations on one consistent basis, and for nothing else.
Written for primary and secondary classroom teachers in state and private schools. University lecturers, private tutors and online course instructors face different, and in the tutoring case much greater, pressure.
What is actually changing#
The unit of analysis is the task, not the job title. A role is not replaced — its task mix shifts.
Preparing lessons and materials
Being augmented✓ Evidence-backedPlanning the sequence, writing worksheets, finding examples, differentiating for three ability levels.
Generating worked examples, quiz variants and differentiated versions of a text is fast and good enough to use, and teachers were among the earliest heavy adopters of the general-purpose tools. The planning judgement — what this class needs next given what happened yesterday — still comes from the person who was in the room yesterday.
Marking and written feedback
Automating≈ Platform inferenceGrading a stack of essays or problem sets and writing a comment on each that the student will actually read.
First-pass marking of structured work and draft feedback on writing are among the most-used education applications, because they attack the task teachers most resent. The teacher still moderates the grade and decides what to say to this child; the hours of routine reading shrink substantially.
Automated feedback on writing is only as good as the rubric, and students learn quickly to write for the machine. Schools that let it replace rather than assist the teacher's reading have seen the quality of feedback fall.
Running the room
Still human-led≈ Platform inferenceThirty children, one adult, forty minutes: attention, behaviour, safety, and the child who is quietly not okay.
This is a physical, relational task with legal responsibility for minors attached, and no technology under development addresses it. It is also, for most teachers, the majority of the working day, which is why the occupation's overall exposure is low despite its planning and marking tasks changing fast.
Explaining, and noticing who did not get it
Being augmented✓ Evidence-backedTeaching the concept, reading the faces, trying a second explanation, deciding whether to move on.
One-to-one tutoring systems explain patiently and adapt to a single learner, and for motivated students with access they are a real complement. Reading a room of thirty, most of whom did not choose to be there, is a different problem: the adaptation is social as much as cognitive, and it is the part parents and systems continue to pay a human for.
Orchestrating tutoring tools
New task✓ Evidence-backedDeciding which tools students use for what, checking what the tools taught them, and catching what they got wrong.
Where schools have brought tutoring systems into class, the teacher's role has shifted towards designing how they are used and monitoring their output. Teachers also now have to teach students how to use these tools honestly, which is new curriculum that lands on them with no extra time.
Which technologies matter here#
Four separate signals. They are deliberately not added together — a job exposed to two technologies is not twice as exposed.
How it got here#
The index is not a static number. This is where it would have sat at each capability checkpoint since ChatGPT — reconstructed, and labelled as such.
● 2 verified events for this occupation, plotted at the date it happened — the parts of the line near a marker are anchored to something checkable.
Starts near the floor and rises steadily, but the rise is confined to planning and marking. The reason the level stays low is that most of a teacher's day is thirty children in one room, which no technology under development addresses. A slow curve here is not a forecast of safety — it is a statement about which tasks have been reached so far.
A flat line is not a forecast of safety. It says which tasks automation has reached so far — the occupations that moved least here are the ones where the constraint is physical or regulatory, and both of those can change.
Recent changes#
One rural district, one stationary robot bought as a teaching aid for a high-school robotics course — not a system that teaches a class. What stopped it was procurement, student-data agreements and union objection, not any limit on what the device can do.
Failure, rollback, regulation or cost is suppressing adoption. Can lower an assessment or widen its uncertainty.
AP News ↗One small country, grades 10–11, a three-year pilot; a Tartu/Stanford/OpenAI impact study is still pending. The student-facing tutor app (Estonian-language, built with OpenAI) only opened in late January 2026, and 47% of activated accounts used it weekly.
Small-scale trial in a real setting. Tells us the deployment conditions are being tested, not that they hold.
AI Leap Foundation (TI-Hüpe) — programme update ↗What this means for you#
Teaching is one of the lower-exposure occupations on this site, and that is a defensible conclusion rather than a comforting one: the classroom is a physical, relational job with legal responsibility, and the tools change the desk work around it. What is genuinely changing for a new teacher is the expectation of tool fluency and the pressure to teach honest use to students. The economics of the profession — pay, class sizes, workload — are a much bigger factor in whether to enter than automation is.
Your exposure is in the hours outside the classroom, and those are the hours you most want back. Used well, the tools return preparation and marking time; the risk is that the system absorbs that time into larger classes or more administration rather than returning it to you. Being the teacher who defines how tools are used in your department is a position worth taking before it is defined for you.
Your options#
Four directions, each with its real constraints and one thing you can test this week. Continuing as you are is a legitimate choice — it just has to be a chosen one.
Take back the desk hours, on purpose
The tools genuinely shorten preparation and marking. Whether that becomes your time or the school's depends on who notices first.
School policies on tool use vary and some forbid it; data protection rules on student work apply.
Time one week of marking with and without a first-pass tool, following your school's rules. Bring the two numbers to your head of department.
Lead how your school uses the tools
Every school needs someone who understands both teaching and the tools well enough to write the policy and train colleagues. That person is usually promoted into it.
More meetings, more administration, and you become the person blamed when a student misuses a tool.
Write a one-page 'how we use tools in this subject' note — allowed, forbidden, how we check. Show it to two colleagues and see whether they would follow it.
Curriculum, assessment or learning design
Designing what is taught and how it is assessed is becoming more complex as tools change what students can do unaided. It uses classroom experience directly and is desk-based.
Fewer roles than classroom positions, and you lose the part of the job many teachers love most.
Redesign one assessment so that it still measures learning if every student has a tutoring tool. If that problem interested you, this path is real.
Common questions#
On the automation question specifically, yes — the classroom is one of the least exposed workplaces we assess, and the tools mostly remove the parts of the job teachers like least. The reasons people leave teaching are pay, workload, class size and respect, and none of those is a technology question. Decide based on those; on the automation axis alone, teaching is a defensible choice.
Method and sources#
- Assessment date
- 2026-09-10
- Basis of the task judgements
- 3 evidence-backed · 2 platform inference · 0 not enough evidence
- Verified events
- 2