The challenge ESL teachers face when they read about AI in education is a career-anxiety challenge dressed up as a pedagogical question. Will AI replace me? Will my school move to AI-only instruction? Is my career running out of road?
The honest answer is more nuanced than either "AI will replace teachers" or "AI changes nothing". AI will reshape the job. The parts of the job AI can do well will get automated. The parts AI can't do will become more valuable. Teachers who lean into the irreducibly-human parts of the work will have more job security, not less, ten years from now.
YapYapGo is a classroom speaking practice tool for ESL teachers, designed around the work that only humans can orchestrate in a real classroom. This post is about which parts of ESL teaching are durable, which are at risk, and what to do about it.What AI does well in language learning
Before discussing replacement, an honest accounting of what AI is genuinely good at:
- Grammar drilling. Patient, consistent, immediate-feedback verb tense and structure practice. AI is better than most humans at this because it doesn't get bored or rushed.
- Vocabulary review. Spaced repetition, definitions, collocations. AI is at least as good as flashcard apps and faster than human review.
- Written feedback. Identifying typos, grammar errors, awkward phrasing in student writing. AI handles this well at scale.
- Lesson preparation. Generating worksheets, question banks, level-graded reading passages. The time-savings are real.
- Pronunciation feedback. Detecting segmental errors and offering targeted practice.
These are valuable. They also represent a particular profile of work - patient, repetitive, individualised feedback - that AI is naturally well suited to. The teaching profession has been doing this work because no other option existed. AI offers a better option for these tasks. Outsource them.
What AI doesn't do well
The parts of language teaching AI structurally struggles with:
- Classroom orchestration. Managing the energy and pacing of a real room of 30 humans. The micro-decisions about when to push, when to ease, when to switch activities.
- Pair-work facilitation. Setting up productive conversations between students, intervening when they go off-track, scaffolding less confident speakers.
- Social and emotional support. Noticing the student having a bad day, the student who's quiet because they're being bullied, the student who's stuck because they're going through something.
- Intercultural mediation. Bridging cultural register gaps when an idiom doesn't translate, when humour misfires, when politeness norms collide.
- Listening for acquisition moments. The kind of listening we covered in why classmates can listen better than AI - noticing the half-formed word the student is reaching for and supplying it.
- Modelling accent variety, slang, and current idiom. Covered in detail in the accents post and the slang and idioms post.
- Real-time judgment. Deciding to skip the next activity because the class is fried. Deciding to pull one student aside. Deciding the lesson plan is wrong and pivoting.
These are the irreducibly human parts. They're also what good ESL teachers have always done. The shift isn't that the job changes - it's that the proportion of the work spent on these parts goes up as AI absorbs the others.
The replacement question, honestly
Will AI replace ESL teachers? The hardest realistic version of this question:
- A school with 50 ESL students. Today: 3 teachers. In 10 years: maybe 2 teachers plus AI tools, or maybe still 3 teachers doing fewer admin tasks and more classroom time. Both are plausible.
- A self-study learner. Today: pay a private tutor for weekly conversation. In 10 years: AI handles grammar, vocab, written feedback; the human tutor is the conversation partner if the budget allows. The human role narrows but doesn't disappear.
- A school's writing feedback workflow. Today: teachers spend hours marking essays. In 10 years: AI does the first pass, teachers do the second pass on high-stakes work. Teacher time freed for classroom orchestration.
The pattern isn't "teachers replaced". It's "teachers shifted toward the work AI can't do". For most ESL teachers this is genuinely good news because the AI-resistant work is also the more rewarding work.
How to lean into the durable parts
If you're an ESL teacher thinking about your career horizon, the durable skills to invest in:
- Pair-work facilitation. Become the teacher who can get 30 students speaking simultaneously, productively, with real engagement. Tools like the Team Maker, Topic Generator, and Classroom Timer handle the logistics so you can focus on the facilitation craft.
- Classroom energy management. The micro-skill of reading a room and adjusting pacing in real time. AI can't do this. Good teachers do it without thinking.
- Feedback craft. Giving feedback that lands, that motivates, that hits the specific point a student needs - not the generic feedback an AI generates.
- Group dynamics. Building a classroom culture where students will speak in front of each other. This is the Willingness to Communicate work covered in the WTC research post.
- Cultural and intercultural fluency. Especially valuable in international ESL contexts where teachers bridge between students from different L1 backgrounds.
The teachers who lean into these skills will be more valuable, not less, as AI handles the work that surrounds them.
Practical near-term moves
For the next school year, three practical moves any ESL teacher can make:
- Adopt AI for the work it does well. Stop spending hours on grammar correction and vocabulary worksheet generation. Outsource these to AI tools you trust and reclaim the time.
- Spend the freed time on classroom craft. More planning of pair-work activities, more thoughtful feedback, more attention to individual students.
- Audit your weekly lesson plan. What share of your classroom time is currently spent on AI-replaceable work (lectures, drills, written exercises)? What share is on human-only work (pair conversations, group activities, real-time feedback)? Shift the ratio toward the latter.
We've also covered the related question of what speaking English actually means, which is helpful framing for arguing the value of human-led work to school administrators who might otherwise be tempted by AI cost savings.
The bottom line
AI won't replace ESL teachers, but it will reshape the job. The work AI does well (drilling, correction, prep) will get automated. The work AI can't do (orchestration, facilitation, real-time judgment, social presence) will become more valuable. Teachers who lean into the human-only parts will have stronger careers in 2035 than teachers who try to compete with AI on the work AI is better at. The future-proof move is to do more of what only you can do.
Sources:
- Selwyn, N. (2022). Should Robots Replace Teachers? AI and the Future of Education. Polity.
- Holmes, W., Bialik, M., & Fadel, C. (2019). Artificial Intelligence in Education. Center for Curriculum Redesign.