Opens in a new tab
needsahuman.

Will AI replace substitute teachers, short-term?

A little.

Most of the day is supervising and managing someone else's class in person, which tools can support but not take over. This job scores 78 out of 100 on (higher is safer). Today people do 10% of the work with AI’s help, and 90% still needs a person.

Updated 3 October 2026 25-3031 2026-Q4
Educational Instruction and LibrarySubstitute Teachers, Short-Term25-3031 · 2026-Q4
0% AI does it10% AI helps90% needs a human
Your job's name, lit by the work that still needs a human.Needs a human 90%AI helps 10%AI does it 0%

AI does it: AI can do the task largely by itself. AI helps: a person still does it, faster with AI. Needs a human: AI can do little of it yet.

Why schools still call a person

A short-term substitute walks into a room of students who did not pick them, on a day nobody planned for. The first job is not instruction. It is taking attendance, learning names fast, and keeping order with a group that knows the regular teacher is gone. That is live judgment about real children in a real room, minute by minute.

The second job is carrying out the plan the absent teacher left, and fixing it when it does not fit. Plans run short. A lab has the wrong materials. A student melts down. The sub has to adapt on the spot, then hand back a usable note on what got done. Software can draft a worksheet in seconds. It cannot read a room, break up an argument in the hallway, or decide that today the class needs quiet reading instead of a group task.

There is also a legal and custodial side. Schools need a named adult responsible for a class, for fire drills, for lunch and recess duty, for the medical note that says one child cannot eat peanuts. That responsibility sits with a person by design, not by habit. It is the main reason the question of whether AI will replace substitute teachers keeps coming back with the same answer: the tools shift paperwork, not custody.

What AI does, what it helps with, and what it leaves to people

The narrowest group is the work AI can take end to end. Marking short quizzes with clear right answers, and turning a thin set of instructions into a printable activity, are the obvious candidates. Both are prep and cleanup, not class time. The share of task time in that group prints in the split above: 0%.

A bigger group is assisted work. Writing the end-of-day note for the returning teacher goes faster with a draft. So does pulling a backup activity when the lesson finishes ten minutes early, or rewording an explanation for a student who did not get it the first time. The person still decides what to use and what to bin. That assisted share also prints above: 10%.

The rest stays with the adult in the room: maintaining order and discipline, supervising students in halls, the cafeteria and the yard, and reading when a quiet kid is actually in trouble. Our Can AI do it? figure for this job, explained on the coverage method page, is 17 out of 100 (higher means more of the task time is already doable). The human-only share sits at 90%.

What the evidence actually shows

There is no published head-to-head test of an AI system against a short-term substitute covering a real class. That is why the evidence grade on this page is D, and why we give no Is it better than a person? number here. A grade like that means not measured, not measured and failed. How the grades work is set out on the quality parity method page.

What would settle it is specific: a study that puts an AI-led classroom session next to a substitute-led one, with the same students, and measures behavior incidents, time on task and what the returning teacher finds the next morning. Tutoring trials do not answer it, because one-to-one tutoring is a different job from covering 28 students for a day.

The labor data is firmer. The Bureau of Labor Statistics counted 524,770 short-term substitute teachers in the United States, with median pay of $41,670 a year, and projects employment up about 2% from 2025 to 2035 (BLS, 2025). That is slow growth, not contraction. Teacher absence does not fall because a model writes better lesson plans.

When this could change

Most likely between 2035 and 2060 (8 in 10 of our scenarios). What that window measures is explained on the replacement year method page.

Two things could pull it earlier. First, a push toward remote or hybrid coverage, where one certified adult supervises several rooms with software running the instruction. Second, tighter school budgets: the cost panel above shows AI tooling landing far below a staffed day, and that gap is what makes districts experiment.

Two things hold it back. Physical presence is a large part of the role, and the robotics tier on this page is dexterous humanoid, which is not something schools can order today. And supervision rules, background checks and parental expectations move slowly. A district that drops the adult in the room is the district that owns what happens next.

What to do: if you sub regularly, ask your district what its AI classroom policy says, because being the person who understands the tool is a reason to be called back.

How to stay needed

Lean into the parts that stay with people. Classroom management is the first: arriving with a routine, setting expectations in the first five minutes, and handling conflict without escalating it. Supervision duty is the second, including the quiet judgment calls in halls and at lunch. The third is adapting instruction live, so the class does not stall when the left-behind plan falls apart.

Two skills raise your value. One is using AI tools well for prep, so you arrive with backup activities and leave a clear handover note. The other is behavior and de-escalation training, including support for students with IEPs, which districts struggle to cover and rarely trust to a stranger.

If you want a longer runway, look at the work next door: Elementary School Teachers, Teaching Assistants and Tutors all use the same room skills with different pay and exposure. You can put any two of them side by side on our job comparison tool, see the wider other teachers and instructors family, or read how school roles score together. Our scoring is open on the methodology page, and the entry-level hiring tracker shows where the first rungs are getting thinner.

Frequently asked questions

Will artificial intelligence replace teachers?

No credible labor projection shows AI taking over teaching as a whole job. What is changing is the task mix: planning, grading and feedback drafts move toward software, while class time, supervision and student relationships stay with people. The Bureau of Labor Statistics still projects growth for most K-12 teaching roles through 2035 (BLS, 2025). The task list above shows which parts of substitute work already lean on tools.

What are the disadvantages of AI teachers?

An AI system cannot take legal responsibility for a room of minors, run a fire drill, or step between two students. It does not notice a child who is suddenly withdrawn. It can state wrong facts confidently, and it has no record of what a specific class did last week unless someone feeds it that. It also cannot enforce rules it has no authority to enforce.

Are computers better at teaching than human teachers?

On narrow, well-defined practice, software can drill and give instant feedback at a scale no person matches. On motivation, behavior and reading a confused face, people are still ahead, and no published study has tested an AI system against a substitute covering a live class. The evidence section above explains what kind of trial would answer this properly.

Will substitute teachers be replaced by AI in 10 years?

Our estimated window for this job, with its full range, is in the chart above rather than typed here. The short version: the physical and custodial parts of the role are the bottleneck, not the lesson content. Districts would need remote supervision models, changed staffing rules and parent acceptance before a day of coverage ran without an adult present.

Are schools already using AI to cover absent teachers?

Mostly for the paperwork around coverage, not the coverage itself. Scheduling systems match subs to assignments, and plenty of substitutes use chat tools to prep a backup activity or write the end-of-day note. Some districts pilot software-led independent work for older students with one supervising adult, which reduces headcount in a building rather than removing the adult entirely.

Is substitute teaching still worth starting?

It remains one of the most accessible ways into school work, usually needing a bachelor’s degree or less depending on the state, plus a background check. Median pay is modest at $41,670 a year for full-time equivalent work (BLS, 2025), and shifts are irregular. Many people use it as a route into a full teaching license or a classroom support role.

Each ridge is a slice of the job's task time.Needs a human 90%AI helps 10%AI does it 0%
The job’s mark

No two jobs leave the same print

Every job gets its own fingerprint, drawn from its code. The amber ridges are the share of task time that still needs a person. Below them, the same ridges are written out in ones and zeros: slate for the work AI helps with, white for the work AI can do.

Substitute Teachers, Short-Term, O*NET-SOC 25-3031. 90% of the job’s task time still needs a human, so 90 of every 100 ridges are amber; slate is what AI helps with, white what AI can do.

What AI can and cannot do

The tasks that make up the job, from , and where AI stands on each today: , (a person does it, with AI speeding it up) or . 90% of the still needs a human.

Each block is one task; its height is its share of working time.Needs a human 90%AI helps 10%AI does it 0%
The job's task list: the parts AI can do are blacked out.Needs a human 90%AI helps 10%AI does it 0%
Enforce school and class rules to maintain order in the classroom.Needs a human
Answer students' questions.Needs a human
Follow lesson plans designed by absent teachers.Needs a human
Take class attendance and maintain attendance records.AI helps
Supervise students during activities outside the classroom, such as recess, lunch, and field trips.Needs a human
Teach social skills to students, such as communication, conflict resolution, and etiquette.Needs a human
Tutor or assist students individually or in small groups.Needs a human
Distribute or collect tests or homework assignments.Needs a human
Distribute teaching materials, such as textbooks, workbooks, papers, and pencils, to students.Needs a human
Teach a variety of subjects, such as English, mathematics, and social studies.Needs a human
Operate equipment such as computers or audio-visual aids to supplement presentations.Needs a human
Counsel students with adjustment or academic problems.Needs a human
Attend professional meetings, educational conferences, or teacher training workshops to improve professional competence.Needs a human
Provide students with disabilities with assistive devices, supportive technology, or assistance accessing facilities, such as restrooms.Needs a human
Organize and supervise games or other recreational activities.Needs a human
Assist students with boarding or exiting school buses.Needs a human
Grade students' assignments and exams.AI helps
Restock teaching materials or supplies.Needs a human

Is it better than a person? The evidence

No direct test against people in this job yet. Every study is , and vendor studies are labelled as such.

When could it be replaced?

When AI could largely do this job: 2035–2060

Most likely between 2035 and 2060 (8 in 10 of our scenarios). A range from our of how fast AI improves, how fast employers take it up and what holds it back, not a forecast that the job ends. “” has a strict meaning here. Today’s answer is at the top of the page; this is how it could change.

The sand is the human working years left, measured in the same 40-year glass for every job, so a safe trade starts nearly full and an exposed job with a thin layer.

The sand is the human working years left, in the same 40-year glass for every job.Years still needing a humanYears run out

How this job could shift, year by year

Where the job could sit on our scale each year to 2060, across the ten behind its .

Today
Will AI replace this job?
A little.
By 2045
60%
of our scenarios have AI largely doing this job by 2045 (Largely.)
0% still have it mostly needing a person (A little. or Nah.)
By 2060
100%
of our scenarios have AI largely doing this job by 2060 (Largely.)
0% still have it mostly needing a person (A little. or Nah.)

We run this job as ten scenarios spread across its replacement range. In each, the score moves towards the bottom band (Largely: AI could largely do the job) by the year that scenario reaches it, slowly at first and faster later, as adoption usually goes. Each bar splits the ten by the band they put the job in. The model stops at 2060. How the timeline works

Share of this job's scenarios in each verdict band, today to 20600%25%50%75%100%2026: 100.0% of scenarios: AI could do a little of this job (A little.)100%Today2030: 80.0% of scenarios: AI could do a little of this job (A little.)80%2030: 20.0% of scenarios: AI could partly do this job (Partly.)20%20302035: 30.0% of scenarios: AI could do a little of this job (A little.)30%2035: 20.0% of scenarios: AI could partly do this job (Partly.)20%2035: 30.0% of scenarios: AI could mostly do this job (Mostly.)30%2035: 20.0% of scenarios: AI could largely do this job (Largely.)20%20352040: 30.0% of scenarios: AI could partly do this job (Partly.)30%2040: 20.0% of scenarios: AI could mostly do this job (Mostly.)20%2040: 50.0% of scenarios: AI could largely do this job (Largely.)50%20402045: 10.0% of scenarios: AI could partly do this job (Partly.)10%2045: 30.0% of scenarios: AI could mostly do this job (Mostly.)30%2045: 60.0% of scenarios: AI could largely do this job (Largely.)60%20452050: 20.0% of scenarios: AI could mostly do this job (Mostly.)20%2050: 80.0% of scenarios: AI could largely do this job (Largely.)80%20502055: 10.0% of scenarios: AI could mostly do this job (Mostly.)10%2055: 90.0% of scenarios: AI could largely do this job (Largely.)90%20552060: 100.0% of scenarios: AI could largely do this job (Largely.)100%2060
Will AI replace the job?Largely.Mostly.Partly.A little.Nah.
Share of this job's scenarios in each band, year by year. Updated with every release.
Show the data
YearLargelyMostlyPartlyA littleNah
Today (2026)0.0%0.0%0.0%100.0%0.0%
20300.0%0.0%20.0%80.0%0.0%
203520.0%30.0%20.0%30.0%0.0%
204050.0%20.0%30.0%0.0%0.0%
204560.0%30.0%10.0%0.0%0.0%
205080.0%20.0%0.0%0.0%0.0%
205590.0%10.0%0.0%0.0%0.0%
2060100.0%0.0%0.0%0.0%0.0%

What’s stopping AI taking over?

The things that keep this work with people, strongest first. Each is scored 0 to 100 from work context, licensing and the evidence we have.

Evidence gapNo study yet compares AI with people doing this job, so employers have no proof it is good enough.
Clients want a personFace-to-face contact is rated 4.1 and physical closeness 2.9 out of 5; caring for or serving people is 3.7 out of 5 in importance.
RegulationWorkers rate responsibility for others' health and safety 3.6 out of 5; the sector has its own rules on who may do the work.
LiabilityMistakes are rated 1.3 out of 5 for consequence and decisions 2.8 out of 5 for impact; someone has to answer for them.
LicensingUsual entry requirement (BLS): bachelor's degree.
Physical work36% of the task time is physical; robots have been shown on 70% of that time.

What would it cost to hand the work to AI?

The share of the year AI could handle (352 of 2,080 hours a year), priced two ways. Both are ranges, not quotes.

AI model usage, a year
$40–$3,520
A person’s wage for the same hours
$4,830–$11,120

AI cost covers model usage only: no integration, licences, oversight or the human time still needed to review the work. Human cost is the wage for the same hours, without benefits or overheads. As of 2026-10.

Robots and humanoids

AI software can only take the work at a screen. The rest needs a robot that can do it.

36%
of the task time is physical work
Dexterous humanoid
the kind of robot the physical work would need
Not commercial: no cited robot does most of this work; humanoids are at demonstration and pilot stage.

Source: Anthropic Economic Index, 'What work can robots do?' (30 September 2026); O*NET 31.0 task weights.

Which AI skills does this job lean on?

The job’s task time split by what an AI model would need to be good at, and where models stand today.

Each star is a task, grouped by the AI skill it leans on.Needs a human 90%AI helps 10%AI does it 0%
Writing · 0% of time
Strong
Drafts, edits and translates most routine documents at professional quality.
Analysis · 5.7% of time
Strong
Reliable on structured data and rules; uneven on judgement calls with thin information.
Coding · 0% of time
Strong
Agents complete many routine software tasks end to end; larger systems still need people.
Vision and design · 0% of time
Good
Reads documents, images and layouts well; specialist imaging needs dedicated, approved tools.
Speech · 7.8% of time
Good
Voice agents handle routine calls and live interpreting; complex or sensitive calls still go to people.
Planning and agents · 7% of time
Emerging
Multi-step agents work in narrow, well-tooled workflows; open-ended coordination is unreliable.
Physical manipulation · 29.9% of time
Early
Robots handle structured, repetitive handling; general dexterity outside fixed settings is not commercial.
Care and persuasion · 49.5% of time
Limited
Can script, coach and advise; trust, presence and accountability still need a person.
Amber matter holds its orbit, slate circles the inner disc, white falls in.Needs a human 90%AI helps 10%AI does it 0%
How exposed is it?

Still needs a human: 78/100↑ safer

The amber matter on the outside holds its orbit: that is the task time that still needs a person. The slate matter circling the inner disc is the work AI helps with. The white matter is the share AI can do; it spirals in and is gone over the edge.

The matter follows the job’s task split: 90% needs a human, 10% AI helps, 0% AI does it. Still needs a human: 78/100 ↑ safer. Will AI replace them? A little.

People are asking

How often people ask whether AI will replace this job: on Google, and by estimate, in AI assistants.

In the US

Under 10
Google searches a month, 12-month average to
1
estimated questions to AI assistants in September 2026
Estimated questions to AI assistants a month, October 2025 to September 2026: from 0 to 1

Source: DataForSEO, US and UK, fetched October 3, 2026. Google figures are Google Ads’ rounded monthly averages. The AI figure is DataForSEO’s estimate from Google’s “People also ask” data, not a count from any AI assistant. UK workers are ONS employment figures matched to this job, so the UK rate per 1,000 is an estimate. Search figures are not part of our open dataset.

What the AIs say

We asked four AI assistants the same question: will AI replace this job in the next 10 years? One word (Yes, Partly or No), then one sentence. Our own answer, from the data: Still needs a human: 78/100 ↑ safer. Will AI replace them? A little.

ChatGPTPartly

AI may handle some supervision, tutoring, and lesson-support tasks, but schools will still need human substitute teachers for classroom management, safety, and student support.

gpt-5.5-2026-04-23 · asked 2026-10-03
ClaudePartly

AI could handle routine supervision or basic instructional tasks in some settings, but the need for classroom management, human connection, and adaptability with students—especially younger ones—means full replacement is unlikely within a decade.

claude-sonnet-5 · asked 2026-10-03
GeminiPartly

While AI will handle instructional delivery and administrative tasks, human supervision will still be required for classroom management, student safety, and emotional support.

gemini-3.8-flash · asked 2026-10-03
PerplexityPartly

AI will automate some substitute-teacher tasks, but human supervision, classroom management, safeguarding, and judgment will likely remain necessary.

sonar · asked 2026-10-03

Assistants answer from what they learned in training (Perplexity also searches the web), so they can be confidently wrong, and the same question can get a different answer tomorrow. Our score is built from task data and graded evidence. Answers collected through DataForSEO.

Cite this page

NeedsAHuman.com (2026). Will AI replace Substitute Teachers, Short-Term? A little. Still needs a human: 78/100, higher is safer; release 2026-Q4. https://needsahuman.com/jobs/substitute-teachers-short-term/ (accessed 4 October 2026).

Scores change with each , so cite the release. The data is open under : credit NeedsAHuman.com with a link. Open data · Press

Put the badge on your site

The badge updates itself with each release and links back to this page.

Sources

  • Tasks and work context: 31.0, ().
  • Jobs, pay and projections: US , and 2025–35.
  • How AI is used today: ; Microsoft Research, .
  • What AI can do: our task ratings ( r1) and the quality evidence register.
  • UK names and employment: coding index and .

How each score is built: methodology. Every figure on this page: open data. Release 2026-Q4.