Teachers are already using AI. The question is how.
At some NSW secondary schools, 60 to 70% of staff already use AI tools for planning. They are copying lesson ideas from ChatGPT, asking Copilot to draft activities, and using whatever is fastest to get through the Sunday night planning session.
This is not a trend that needs to be started. It is a trend that needs to be directed.
The problem with generic AI
ChatGPT is genuinely useful. But it has three fundamental limitations when it comes to lesson planning:
It does not know the curriculum. When you ask ChatGPT for a Year 9 Commerce lesson, it will generate something that looks plausible. But the outcome codes are often invented. The content may not align to the NSW syllabus. There is no verification layer.
It does not know your students. A generic lesson plan is written for a generic class. It cannot differentiate for the student with ADHD in the front row, the EAL learner who needs visual scaffolding, or the advanced student who needs extension work — unless you manually prompt for each one, every time.
It does not plan in sequence. Each lesson is generated in isolation. Week 3 does not know what Week 2 covered. There is no term scope, no progression, no coherent sequence of skill development.
The result: teachers spend almost as long editing and verifying AI-generated content as they would have spent planning from scratch.
What purpose-built AI looks like
A planning tool built specifically for teachers solves these problems by design:
Curriculum verification — Every outcome code is checked against the NESA syllabus database before it reaches the plan. No hallucinated codes. No surface-level alignment.
Scope before lesson — The term plan is built first. Each week knows what came before it and what comes next. Lessons sit inside a coherent sequence.
Learner profiles — Teachers build anonymised profiles once per class. Every lesson then includes specific teaching adjustments for individual students — woven into every instructional phase, not appended as an afterthought.
Structured output — Table-formatted, timed, exportable to Word. Matches school submission standards. A substitute teacher could pick it up and deliver it.
What school leaders should look for
When evaluating AI planning tools for your school, ask these questions:
- •Does it verify outcome codes against the actual syllabus, or just generate plausible-sounding ones?
- •Does it build a term scope first, or just generate standalone lessons?
- •Does it know the students in the class, or just produce generic differentiation advice?
- •Does it produce structured, editable output that meets your school's submission standards?
- •Is student data anonymised and handled appropriately?
If the answer to any of these is no, you are looking at a content generator, not a planning system.
The leadership opportunity
AI in lesson planning is not going away. Teachers are already using it. The opportunity for school leaders is to direct that energy toward tools that actually meet curriculum standards, reduce workload meaningfully, and maintain the professional judgement that makes great teaching possible.
Done well, this is not about replacing teachers. It is about removing the scaffolding work that sits between a teacher and a great lesson — and giving them their evenings back.
Smart Syllabus Assistant is purpose-built for NSW secondary teachers. Scope and sequence through to differentiated lesson plans, all aligned to NESA outcomes. See how it works or start free.