IGNITE English Resources

Course Design

It's Never Too Late to Redesign the Course

Most of the assignments in your courses were built to solve a problem that no longer looks the way it did. That's not a criticism of the assignments. It's a description of what happened to the world in between.

You Already Redesign Courses. This Is the Same Job.

Ministry of Education documents in Ontario are unusually direct about this in writing. Growing Success defines professional judgement as a purposeful and systematic thinking process that evolves in terms of accuracy and insight with ongoing reflection and self-correction. Evolves. Self-correction. Those words sit in the policy that governs how we as educators assess, and they describe a practice that keeps moving, rather than one that gets set down once and defended afterwards.

You've done this many times without calling it curriculum redesign. You changed how you taught paragraph structure when a cohort couldn't figure out how to write an effective thesis statement. You dropped texts that stopped resonating with your students and found ones that did. You rewrote a unit after a semester where half the class arrived reading two grade levels below where the course assumed. Nobody framed those changes as reform. They were just the job.

Redesigning for AI is the same move, prompted by a change that happens to be larger and faster than most. It doesn't require you to become a different teacher. It requires you to look at what you already have and ask whether it still demonstrates what you think it demonstrates.

What Changed Isn't the Thinking. It's the Evidence.

This is the part worth being precise about, because it's where most of the panic comes from.

What you want from a Grade 10 student hasn't changed. You still want them to read closely, hold an argument across a piece of writing, choose evidence that earns its place, and say something that sounds like a person rather than a generic summary. None of that is obsolete and none of it needs rewriting.

What changed is that the artefact you used to collect as proof of that thinking can now be produced without the thinking. The essay itself was never the complete point. It was the evidence showing how they got there. For about thirty years it was reliable evidence, so we stopped noticing the difference between the two. That distinction is the whole redesign.

Ask of every task you already use: if this arrived finished and correct, would I actually know if the student did the thinking? If the answer is no, then the problem isn't the student. It's that the task collects, and values, the wrong thing.

The Freeze Is Understandable. Waiting Is Still a Choice.

It's worth saying plainly that most teachers are being asked to handle this without much help. RAND found in spring 2025 that only about a third of teachers reported having a school or district policy on AI and academic integrity, while more than half of English, Mathematics and Science teachers had used AI themselves that year. Practice is well ahead of guidance, which is an uncomfortable place to work and isn't a personal failing.

But the absence of a policy doesn't pause the problem. Pew surveyed 1,458 American teenagers in the autumn of 2025 and found that 54 percent had used a chatbot for help with schoolwork, with about one in ten saying they do all or most of their schoolwork that way. In the same survey, 59 percent said they believed students at their school use AI to cheat at least somewhat often. Those are American figures and Canadian classrooms may differ at the margins, but not by enough to plan around.

So the choice isn't between changing and staying still. Staying still is a decision to keep marking work you can't vouch for, and to keep having the difficult conversation at the end rather than designing it out at the start.

You Can't Design Against a Tool You've Never Used

The argument for learning this yourself isn't a moral one. It's practical. If you haven't spent an hour with a chatbot and your own assignment sheet in front of you, you'll guess wrong about what it can and cannot do.

Teachers who haven't tried it tend to overestimate it in some places, assuming it can produce specificity it can't, and underestimate it badly in others, assuming a creative or personal task is completely safe when it's among the easiest things to generate. You can't build a defence on a guess about the thing you're defending against.

There's also a straightforwardly selfish reason. In a Walton Family Foundation and Gallup survey of 2,232 American public school teachers in spring 2025, those who used AI weekly reported saving 5.9 hours a week on planning, marking and administrative work. That's a real amount of time, and the honest use of some of it is the redesign this article is asking you to do. Learning the tool is what buys you the hours to fix the assignment.

Most of Your Course Doesn't Need to Change

Redesign sounds like starting over, which is exactly why people put it off. That's not what's being asked of you.

Your text choices are fine. Your subject knowledge is fine. Your sense of what a good paragraph looks like is fine, and better than any tool. The learning goals in the curriculum didn't change when the technology did.

What needs to change is narrower than it sounds: how much of the work you see being done in class, what you collect along the way, and whether the task depends on anything only this student could supply. On most assignments that's a handful of edits rather than a rebuild.

Start With One Assignment, Not the Whole Course

If you try to audit a whole course you won't finish, and you'll feel worse than when you started. Take one task instead.

Two different questions get confused here, so it's worth separating them. To find out where you actually stand, audit an ordinary assignment: not your best one, and not the one that worries you most. A typical task tells you something you didn't already know, which is the entire value of a diagnostic. If you test the assignment you're already anxious about, you get confirmation rather than information.

Then, when it comes to spending your redesign time, start with the one that worries you most. That's usually the major take-home essay. Diagnose on the ordinary one, rebuild the frightening one first.

Resistant And Intelligent, Not One or the Other

There are two jobs here, and departments often pick one and argue about it.

AI-Resistant design asks what a machine can't do and builds the task so a shortcut fails. That protects your ability to assess independent thinking, which you still need to be able to do and still must report on.

AI-Intelligent design asks the newer question: what can a student do well now because the tool exists, and how would you assess the quality of that judgement? It permits AI use openly, asks students to declare it and to question how they used it, and they can even hand you the prompts. Then you mark how well they did that.

Most English courses need both, and the same course can hold both without contradiction. One task protects the thing you've always assessed. The next one teaches a skill your students will need whether or not you approve of the tool. Deciding which of the two jobs a given assignment is doing is most of the clarity a department is missing.

Where to Start

Examine one of the courses you're teaching. Not to rebuild it, and not to write a policy. Take an ordinary assignment, run it through the Audit, and change the weakest part of it. That's a single evening, and it's the difference between a course designed for a different issue and one designed for this one.

Questions Teachers Ask

Is it too late to change an assignment I've already handed out?

No. You can add a stage to an assignment already running: an in-class checkpoint, a short conference, or a reflection attached to the submission. Students rarely object to a stage that's clearly about showing their work rather than catching them out.

I teach three, or four, or five courses. Where do I realistically start?

One assignment in one course. Audit an ordinary one to see where you stand, then rebuild the one you'd least like to defend. Fixing the weakest pillar on a single task teaches you the move, and that move transfers. Three courses redesigned badly are worse than one task redesigned properly.

Do I have to allow AI use to do this?

No. The redesign is about what evidence you collect, not about what you permit. A course that bans AI entirely still needs process, in-class work and personal specificity, because a ban you can't enforce isn't a design.

What if my department doesn't agree with me?

Then start with your own courses. A department position is easier to reach once somebody can say what they changed, what it cost them, and what happened. One teacher with a redesigned assignment is a more persuasive argument than a meeting.

Where the Numbers Come From

All of these are real and checkable. The dates and sample sizes are given so you can decide how much weight each one carries, and so you can look them up.

I was unable to find a Canadian survey of comparable size and recency that I could verify, and I'd rather cite real American data than a Canadian number I can't stand behind. That gap is itself worth noticing: if you're reading this and thinking about assessment redesign, you're ahead of the Canadian curve.

Audit One Assignment

An ordinary one, not your best and not the one that worries you most. Eight questions, five minutes, and the result names what to change first.

Get the Free AuditRead the Framework
George O'Toole

About the Author

George O'Toole is a retired Head of Department with over 35 years experience in secondary English classrooms, a former Educational Technology Coach, and a former instructor with the Queen's University Faculty of Education. He holds a Master of Education in Language and Literacy from the University of Toronto and writes from Whitby, Ontario.

Classroom resources on Teachers Pay Teachers