Every few months, a think piece appears claiming that AI will make teachers redundant. Every few months, teachers roll their eyes and keep planning their lessons. Both responses are understandable — and both miss something important.
The more interesting question isn't whether AI will replace teachers. It's whether AI will replace lesson planning as a task — the Sunday-evening hours, the blank-document anxiety, the scheme-of-work construction. And the honest answer is: partly, and sooner than most teachers expect.
What AI can already do well
Current AI lesson planning tools can generate a complete lesson structure — objectives, starter, main activity, differentiation, plenary, homework — for any subject, any year group, in under a minute. For a competent teacher, the output is usually at least 70% usable without editing. That's a genuine time saving.
They can also generate supporting materials: vocabulary lists, quiz questions, worksheet prompts, exit ticket questions. Tasks that would take thirty minutes can take thirty seconds. The mechanical work of lesson construction has already been substantially automated.
What AI cannot do
AI doesn't know your class. It doesn't know that Year 10C has been unsettled since the friendship group split, or that the boy who sits near the window is reading two years below his age and needs a different entry point. It doesn't know what actually happened in the lesson before this one, or what misconception came up in the marking that needs addressing first.
This is structural, not a matter of AI not being good enough yet. Lesson planning isn't a document-creation task. It's a thinking task, and the thinking is grounded in specific knowledge of specific people that no tool has access to unless the teacher provides it.
AI also can't hold the room. The relationship between a teacher and a class — the accumulated trust, the shared references, the calibrated challenge — is what makes a lesson plan come alive. No plan, however well-crafted, teaches itself.
What AI will change
The trajectory is fairly clear: AI will increasingly handle the structural and mechanical parts of lesson planning. First drafts, resource generation, differentiated versions of the same task, unit-level sequencing suggestions. Teachers will spend less time creating from scratch and more time editing, adapting, and applying professional judgement.
This is already happening. The teachers using AI planning tools today are not producing worse lessons — they're producing comparable lessons in a fraction of the time, and spending the time saved on marking, feedback, and students. The cognitive centre of gravity is shifting from production to curation.
The irreducible human element
There are things in teaching that resist automation not because AI isn't capable enough, but because they require a relationship. Deciding that a student needs encouragement rather than challenge today. Recognising that the class needs fifteen minutes of discussion rather than the planned activity. Knowing when to abandon the plan entirely because something more important has come up.
These are teaching decisions, not planning decisions. And they'll remain human for as long as teaching involves humans — which, for the foreseeable future, is all of it.
The honest answer
AI will not replace lesson planning. It will replace the most mechanical parts of it — the blank-page problem, the formatting, the resource creation — and give teachers back the time those parts consumed. What remains will be smaller in volume and higher in value: the professional judgement, the relational awareness, the contextual thinking that only a teacher can do.
That's not a threat to the profession. It's a reasonable description of what good teaching already looks like — just with fewer Sunday evenings lost to paperwork.
Try ScholaPlan free for 30 days
AI-generated lesson plans for any subject, any year group. No credit card required.