AI for ESL Writing – Literacy Tool or Crutch? Three Months of Watching
I teach ESL and English at the secondary level and I’ve been watching how my ESL students use AI writing tools for three months now. My honest take.
For students at higher proficiency levels: AI assistance is behaving like a more responsive version of a writing tutor. They use it to check if their sentence sounds natural, get suggestions, then make their own judgment. They’re engaging with the feedback, not just accepting it.
For students at lower proficiency levels: this is where I worry. They’re accepting AI suggestions without understanding why the change was made. The writing improves but the language acquisition doesn’t. it’s Grammarly on steroids – fast, fluent output without the underlying learning.
The variable that matters most is whether the student can explain the changes that were made. that’s the actual test of whether the tool is supporting or replacing the learning.
What are ESL teachers doing to maintain that line in practice?
6 Replies
Join the discussion.
Log In to Reply
this is the same conversation we have about Grammarly in bilingual schools. Grammarly used as a proofreading tool by a competent writer: fine. Grammarly used to generate sentences by someone who can't write them: circumventing the learning. The distinction isn't the tool, it's the role it's playing.
the "can you explain the change" spot check is something I've been doing intuitively but never framed this clearly. naming it as a practice makes it easier to build into regular feedback cycles rather than using it only when something feels off.
first year teacher here. this thread is basically the entire professional development I didn't get. thank you.
the crutch problem is just the calculator debate from 1985. didn't resolve cleanly then either.
the "can you explain the change" test is exactly right and it's what I use in my university ESL classes. I call it a spot check - I pick a sentence from the revised draft and ask the student to explain why that version is better. if they can, the tool is working as a learning aid. if they can't, it's working as a replacement.
the proficiency-level split you're describing also tracks with my data. advanced ESL learners use these tools more like native speakers use Grammarly - to catch errors they already suspect. lower proficiency learners use them to generate language they couldn't produce themselves. completely different cognitive processes.
Holly's split between proficiency levels is what I see too. the tool is essentially amplifying existing ability - students who already write well use it like a spell checker, students who struggle use it as a ghostwriter. same tool, completely different cognitive engagement.