Should Teachers Use AI to Plan Lessons? The Case For and Against

The question is what it costs you

Many teachers are not frightened of AI. They are unconvinced by it. They have watched waves of education technology arrive with confident claims, and they are entitled to ask a harder question than “How quickly can it make a worksheet?” What does it cost me to use it?

That cost might be money, time spent checking a draft, or the gradual loss of practice in making decisions you used to make yourself. A product demonstration usually shows the document arriving. It rarely shows the teacher finding the wrong example, rebuilding the sequence or deciding not to use it.

Classmate sells an AI planning tool, so we have a commercial interest in the answer. That is a reason to examine the argument carefully. Here is the case for using AI to plan lessons, the case against, and a way to decide which applies to your work. Deciding that a tool is not useful for you is a reasonable outcome.

Why teachers are wary of AI lesson planning

Will it replace me? A company selling software cannot promise what every school or employer will do. What it can acknowledge is that drafting a plan and teaching a class are different responsibilities. A plan is a document; teaching happens in a room, with students whose understanding, attention and circumstances change. Producing the first does not demonstrate an ability to do the second.

Will colleagues think I am cutting corners? This is a real professional concern. A polished plan does not explain which decisions were yours or whether you checked the output. Be prepared to show what you changed, why the lesson suits your students and how your use fits school expectations. Speed alone is not a defence, and scepticism does not disappear because a supplier calls its tool innovative.

What about students using it? Teacher preparation and student assessment have different purposes. A student may need to demonstrate unaided knowledge in an assessment; a teacher may be using an approved tool to prepare practice material. Those positions can coexist, but the distinction needs explaining rather than hiding. Follow the rules for the actual task, and discuss student use separately with your school.

Will I get worse at planning? This is the concern worth taking most seriously. It is also the one a conversation about convenience can too easily avoid.

Does using AI to plan lessons make you a worse teacher?

It can, if it becomes a way to avoid the decisions through which you develop your teaching. That is a risk mechanism, not a claim that every teacher who uses AI will become less capable, or that research has established a particular rate of decline.

Planning is not administration that happens to precede teaching. It is where a great deal of the thinking happens. When you take a content description and work out how to move thirty students from their current understanding towards it, you make decisions: which misconception to surface, where the lesson may stall, what needs modelling and which example will work for this class.

Accepting a generated sequence without interrogating it skips that practice. An occasional draft is different from a habit of accepting every draft. The concern is what becomes routine when the work keeps arriving and the output keeps looking plausible.

A bad plan is relatively easy to reject. A plausible plan can pass a quick read even when it contains an assumption you would have questioned if you had built the sequence yourself. The risk is therefore not limited to obvious errors. It includes the important thinking that a neat document lets you overlook.

What planning actually teaches you

Some parts of preparation deserve protection because they develop and exercise professional judgement. Deciding what students already know, choosing what has to come before what, estimating realistic lesson time and choosing evidence of understanding are central to the work. A tool can offer suggestions, but you still need a reason for accepting them.

Other parts are document production: applying headings, moving approved wording into a template, reformatting a settled explanation into slides or laying out practice questions that you have already selected. Doing those tasks by hand is not automatically a better use of a teacher's time.

The boundary matters. Writing success criteria is not merely rewording a curriculum description: choosing observable evidence requires judgement. Designing a worksheet can reveal a gap in your explanation. Deciding what belongs on a slide affects what students notice. The safe task to hand over is the production that follows a decision, not the decision concealed inside it.

How much of your preparation falls on each side will vary by subject, experience and lesson. There is no honest universal percentage. The case for AI is that it may reduce avoidable production work while leaving you with enough time and attention to improve the teaching.

The case for AI: a draft worth arguing with

A first draft can be useful when you know enough about the lesson to challenge it. Instead of spending your first ten minutes arranging a blank document, you can compare a proposed explanation with your own, reject an unsuitable example and decide what needs a clearer model.

Turning an agreed plan into a presentation or worksheet is another possible saving. The benefit depends on consistency: if the slides introduce different vocabulary or the worksheet assesses something you never taught, the conversion has created more work. Check the relationship between the resources, not just whether each file looks finished.

AI can also suggest starting points outside your specialism. A teacher covering an unfamiliar topic may welcome alternative explanations or examples. But this is precisely where checking can be harder. A fluent draft should prompt consultation with a subject colleague or authoritative source, not confidence you have not yet earned.

In each case, the advantage is conditional. Count the time spent specifying, reviewing, correcting and preparing the final resource. A fast generation that takes an hour to repair has not saved that hour.

The case against AI: when the checking costs more

AI does not know your class simply because a prompt contains a year level. It has not heard yesterday's discussion or noticed which students can repeat a method without understanding it. Context you supply can improve a draft, but it cannot turn the tool into a colleague who has taught those students.

It can also make a plan sound complete by filling gaps with generic activities. “Discuss in groups” may conceal an unanswered question about what students are ready to discuss. “Differentiate as needed” may conceal the absence of a usable scaffold. An attractive layout does not repair either problem.

Curriculum references, factual examples, quotations and calculations need checking, including in Classmate. If you cannot verify the content confidently, or the tool repeatedly drifts from your brief, using an existing approved resource or planning with a colleague may be the more efficient choice.

There is also a legitimate reason to plan without AI when you want to practise: learning a new syllabus, developing an unfamiliar teaching sequence or working through a misconception yourself. Time spent thinking is not wasted just because a machine could produce a document sooner.

Is AI lesson planning cheating?

Using a drafting tool does not by itself answer whether the work is professionally acceptable. The relevant questions are what your school permits, what the task requires, whether you are transparent where required and whether you can take responsibility for the final lesson.

A routine lesson-preparation task is different from an assessment of your independent planning in initial teacher education or accreditation. Follow the conditions of that task. Do not present generated material as independently produced work when those conditions require otherwise.

The Australian Framework for Generative AI in Schools puts teaching and learning, transparency, accountability and privacy among its guiding principles. It provides a basis for discussing responsible use; it does not grant permission for every product or override your school's rules. See the Australian Government's framework: education.gov.au

The practical ethical test goes beyond disclosure. Would you defend the lesson's decisions, correct an error and explain the source of the material? If the only defence is that the software generated it, responsibility has been left in the wrong place.

Give context without giving away student information

Bring useful teaching context: the prior concept students have learned, the vocabulary the task requires, the equipment available and the amount of lesson time. “The class needs oral rehearsal before independent writing” is more useful than asking for a generic mixed-ability lesson.

Leave out names, student records, diagnoses, photographs and other identifying details unless your school has specifically approved that use with that tool. Removing a name is not necessarily enough if the remaining description identifies the student. You can describe the teaching need without uploading a personal history.

Read the product's data terms and your school's guidance before using real classroom information. A curriculum-planning tool is not automatically an approved place for confidential material. The national framework includes privacy and security as a principle; the local decision about approved tools still matters.

Keep the decisions visible: one worked example

Imagine a class can identify a persuasive claim but struggles to explain how evidence supports it. You decide that the next lesson should practise the link between one claim and one piece of evidence. That decision comes before generation.

An AI draft proposes a debate, a long reading and an independent essay in one period. Each activity could look educational on its own. Together, they may bypass the explanation your students need and exceed the time available.

You replace the debate with a short teacher model, choose a manageable extract and prepare a partially completed explanation for guided practice. You ask students to explain a second example independently as an exit task. AI may help format the slides or draft alternative practice examples, but you verify the text and evidence yourself.

The point is not that this is the right lesson for every class. It is that you can explain why you changed the draft. The sequence follows a diagnosis of the learning need rather than the order in which the tool produced activities.

How to use AI without outsourcing your judgement

Read the draft as a critic, not a recipient. “What would I change, and why?” is a stronger question than “Does this look good?” Check the explanation, the sequence, the pace and what students must think about at each step.

Verify curriculum references against the official wording and the curriculum your school uses. Do not stop at a familiar-looking code. For Australian Curriculum references use australiancurriculum.edu.au ; for NSW use curriculum.nsw.edu.au ; for Victorian Curriculum Version 2.0 use f10.vcaa.vic.edu.au . A general topic match is not the same as a lesson addressing the selected content.

Edit deliberately rather than repeatedly asking for a different answer. Direct editing can make your decisions easier to retain; a targeted request can also be useful when you know exactly what needs changing. Neither method is a guarantee of ownership. The important step is checking what changed and whether the rest of the lesson still fits.

Notice when you have stopped reading properly. Accepting a fourth plan because the first three seemed fine is an understandable shortcut on a busy day. It is also the moment to slow down, use a known resource or ask a colleague to review the part you are unsure about.

Keep a short note after teaching: what you changed, what worked and what you would do differently. That reflection gives the next lesson a better starting point than another generic prompt.

What to check in every AI-generated lesson plan

Before you teach, ask:

• Is the curriculum reference correct, and does its full wording match the learning I actually intend?
• Does the sequence make sense for what these students already know?
• Where is the lesson likely to stall, and have I planned a useful response?
• Does the differentiation offer meaningful access and challenge rather than just easier and harder questions?
• Will the assessment reveal understanding rather than only completion or compliance?
• Could I defend the sequence to a colleague who asked why I chose it?

If you cannot answer the sixth question, you have received a draft. You still have planning to do. If any of the earlier answers is unclear, revise the relevant part before teaching rather than assuming you can improvise around it in the room.

What this means for Classmate

Classmate helps teachers draft and edit lesson plans, presentations and worksheets. That is a practical offer, not a claim that generated material is ready to teach without review. Teachers still choose the curriculum focus, check the content and decide how to use it with their students.

Classmate's lesson plans do not add achievement standards. Any checks your school requires against those standards remain part of your own curriculum planning. Generated resources should be reviewed for accuracy, accessibility and fit, even when they were created from the same lesson brief.

Try a familiar lesson if you want to evaluate the workflow, because you will be better placed to recognise a weak example or missing step. You may conclude that one part saves time and another does not. That is more useful information than whether the first output looked impressive.

So: should teachers use AI to plan lessons?

Yes, if you use it as a draft and remain the person who decides. No, if using it means accepting a sequence you cannot explain, overlooking your school's requirements or spending more time repairing material than preparing it another way.

The deskilling concern should not be dismissed as resistance to change. It points to the distinction that matters: planning develops judgement, while some of the paperwork around planning merely records it. Protect the former. Delegate the latter only when doing so helps.

There is no professional duty to use AI because it exists, and no virtue in doing avoidable formatting by hand. The choice is whether this use, for this lesson, leaves you doing the thinking that planning was always for.