The Problem
Your students are already pasting their emails into ChatGPT and asking it to “make it better.” The AI rewrites everything at once (grammar, tone, word choice, etc.) producing a polished result but zero learning. Students can’t see what was actually wrong, don’t understand why changes were made, and gradually lose confidence in their own second language abilities. Over time, this creates “learned helplessness”: students who can no longer write professionally without AI assistance.
A Better Approach: From Crutch to Learning Tool
Instead of asking ChatGPT to rewrite, teach students to ask it to explain. The goal shifts from getting a better email once to understanding principles they can apply every time they write.
Four-Step Framework
- Students write first (without AI) – Preserves ownership and gives them concrete text to analyze
- Explanation before correction – Ask ChatGPT “what are the issues?” not “rewrite this”
- Keep original visible – Student’s text stays central; explanations exist around it, not instead of it
- Revise selectively – Work on 1-2 issues at a time to prevent “everything I wrote was wrong” feeling
Sample Prompts That Build Understanding
Instead of: “Rewrite this email to sound professional”
Try these:
For tone: “How might this message be interpreted by the reader? Does it convey [intended tone]?”
For grammar: “In the sentence [X], are there any grammatical errors? Please explain them without rewriting.”
For clarity: “What questions might a reader have after reading this sentence? What information seems missing?”
For strengths: “Which parts of this email are already effective and appropriate?”
For L1 transfer: “Does this phrase sound like a direct translation from [L1]? What would be more natural in the second language?”
What Students Learn
When students use this analytical approach instead of asking for rewrites, they develop:
- Transferable grammar knowledge: Specific rules they can apply to all future writing, not just one-time fixes
- Pragmatic awareness: Understanding how word choices affect professional relationships and reader perception
- Tone control: The ability to distinguish between being clear/direct and being harsh/accusatory
- L1 transfer recognition: Identifying when they’re translating directly from their first language instead of using natural second language patterns
Critical AI Literacy
Teach students that ChatGPT isn’t always right. Red flags include:
- Overly prescriptive rules about style preferences
- Cultural assumptions that don’t match your workplace
- Contradictory advice between different prompts
- “Rules” that are actually just one way to phrase things
Three Ways to Start
- Model it: Demonstrate the analytical approach with a sample email in class. Show what good prompts look like.
- Create a prompt library: Build a shared document with effective prompts for common issues (grammar, tone, L1 transfer). Add to it as students discover useful questions.
- Start a conversation: Ask students how they currently use AI. Explore whether it’s building skills or creating dependency. Help them connect AI use to their learning goals.
The Bottom Line
Students will use ChatGPT regardless of what we say. Our role is teaching them to use it as a learning tool rather than a replacement writer. By teaching better questions, we help students remain active writers rather than passive editors of AI-generated text as well as build the critical thinking skills they’ll need in an AI-rich professional world.
This article was written by Ashton N., ESL Instructor at Language Advantage Inc.
