AI for English teachers: what to hand over and what to keep
The question is not whether AI replaces English teachers. It is which parts of the job are worth your evening — and which were never the teaching.
Most of an English teacher's unpaid hours go on production: worksheets, gap-fills, reading texts at the right level, comprehension questions, model answers. None of that is teaching. It is the manufacturing that makes teaching possible, and it is where AI helps most.
Hand over: material production
Levelled reading texts. Ask for a 200-word passage on a named topic at a stated level, using a specific tense. Getting text at the right level was previously the hardest thing to source.
Practice sets. Twenty sentences drilling one structure, with vocabulary from your learners' actual context — Kochi, not London.
Differentiation. The same worksheet at three levels for a mixed class, in one request. This used to be the reason mixed classes got taught to the middle.
Model answers. For an IELTS task or an exam question, plus a version with deliberate errors for learners to correct. That second one is a genuinely good activity and tedious to write by hand.
Rubric-based first-pass marking. Paste your rubric and a script, get an initial breakdown. As a first pass to speed up, not as the final mark.
Keep: everything that requires a person
Deciding what to teach next. That comes from what your class could not do yesterday. No tool knows that.
Correction in the moment. Judging whether to interrupt an error or let fluency run is a live decision, and it is most of skilled teaching. See fluency vs accuracy.
Pronunciation assessment. Automated scores are unreliable for Indian English in both directions — why.
Managing the room. Who is hiding, who is dominating, who understood but will not say so.
The final mark. A grade with consequences needs a person who can defend it.
The failure mode to watch
AI-generated language material has a specific problem: it is fluent, plausible, and sometimes wrong in ways that survive a quick read.
Common examples worth checking every time: a "past simple" exercise with three present perfect sentences in it; a gap-fill where two options both work; a comprehension question whose answer is not in the text; vocabulary at the wrong level despite the level you asked for; and invented example sentences that no native speaker would produce.
The rule that holds: you are still the author. A worksheet you did not read before printing is a worksheet you cannot defend when a learner asks why an answer is right — and being unable to answer that question in front of a class costs you more than the hour you saved.
A prompt worth reusing
Vague requests produce vague material. This shape works:
Write eight gap-fill sentences practising the past simple of irregular verbs, for adult intermediate learners in Kerala. Use everyday contexts — travel, work, family. One correct answer only per gap. Give the answer key separately, and list any verbs above the target level.
The last two clauses do the work: separating the key makes the sheet usable, and asking it to flag over-level vocabulary makes it check its own output.
Teaching is a craft you practise under observation, not one you prompt your way into — which is why the teacher training course and the IETTC are built around live teaching with feedback. AI buys you back the evening. What you do with the class is still yours.
Frequently asked questions
- Will AI replace English teachers?
- It replaces material production — worksheets, levelled texts, practice sets, first-pass marking. It does not replace deciding what to teach next based on what a class could not do, judging when to correct in the moment, assessing pronunciation reliably, or managing a room of people.
- Is AI-generated teaching material reliable?
- Only after you check it. It is fluent and plausible and fails in specific ways: tense exercises containing the wrong tense, gap-fills with two valid answers, comprehension questions whose answer is not in the text. You remain the author of anything you hand to a class.
- How do I get better teaching material from AI?
- Be specific about level, learner context, target structure and quantity, then ask for the answer key separately and for any vocabulary above the target level to be flagged. That last instruction makes the tool check its own output, which catches most level mismatches.
Keep reading
Using AI to practise English speaking: what works, what does not
An AI will talk to you at 6am, never gets bored, and never judges your grammar. It also will not tell you the truth about how you sound unless you ask correctly.
Does VR actually help you learn a language?
VR is sold as the future of language learning. Its real contribution is narrower and more interesting than that — and it is not about the graphics.
Why AI pronunciation scores are unreliable for Indian English
A pronunciation app told you 72%. That number feels precise and is not. Understanding what it measured tells you when to trust it.