Course

Leading AI Adoption in Your Faculty

Roll out AI responsibly — set the policy, bring staff with you, protect students and staff, and measure the impact — while you stay the accountable human leader.

Modules

Each module: a short why, rich content grounded in the Australian Framework and NSW guidance, three reveal-answer knowledge checks, a concrete leadership activity that drafts part of your position or rollout plan, and a facilitator note for running it with a leadership cohort.

Click a module to read it.

1. 1

   The leader's mandate — why this is yours to lead

   Why lead AI now, why the Australian Framework makes accountability yours, and the twin failure modes: the paralysis of doing nothing versus the harm of moving recklessly.~45 min

   By the end of this module you'll be able to:

   • Articulate why faculty AI adoption is a leadership responsibility you cannot delegate or defer.
   • Name the twin failure modes — unmanaged 'shadow' use versus reckless rollout — and the harm each causes.
   • State your own role in plain terms: keep humans accountable and the practice safe.

   Standards7.1 Meet professional ethics and responsibilities6.3 Engage with colleagues and improve practice6.4 Apply professional learning and improve student learning

   This is already happening — the only question is whether you are leading it

   AI is in your faculty now. Staff are rewording feedback, drafting units, generating questions — some of it well, some of it in a public tool with a student's name pasted in. If you have not set a position, you have not avoided the risk; you have left it ungoverned. Leadership here is not optional, and it is not a tech project — it is a duty-of-care and professional-practice responsibility that sits with you.

   The Australian Framework for Generative AI in Schools is explicit on this. Its Accountability principle states that teachers and leaders remain responsible for the decisions AI supports. You cannot outsource the judgement to a vendor, an algorithm, or "the staff who are into that sort of thing". When a parent, your principal, or NESA asks who decided this was safe and appropriate, the answer is a person — and in your faculty, increasingly, that person is you.

   The twin failure modes

   There are two ways to get this wrong, and good leaders can fall into either.

   Failure mode	What it looks like	The harm
   Doing nothing (paralysis)	"We'll wait and see." No position, no guidance, no PD.	Shadow use. Staff use AI anyway — in unapproved tools, with no data discipline, no disclosure norms, no quality control. You carry the risk with none of the control.
   Moving recklessly (hype)	"Everyone must use it." Mandates, no guard-rails, no privacy line, integrity left to chance.	Real harm. Student data in public tools (a privacy breach), hallucinated content taught as fact, assessment integrity undermined, staff deskilled and resentful.

   The job is not to be the most enthusiastic adopter or the most cautious blocker. It is to be the accountable human in the middle — moving deliberately, with guard-rails, bringing people with you.

   What "leading it" actually means

   Strip away the noise and your mandate is small enough to hold in your head:

   • Keep humans accountable. A teacher signs off on every output that reaches a student. AI drafts; a professional decides. You make that the norm, and you model it.
   • Keep the practice safe. Hold the student-data hard line, align to the Australian Framework and DoE guidance, and make sure people are supported — not just told.
   • Keep it honest. Be transparent with your staff, your students and your community about how and when AI is used. Transparency is one of the six principles, not a courtesy.

   Lead the practice, not the product

   A trap worth naming early: this is not about picking the cleverest tool. Tools change every term. What you are actually leading is a change in professional practice — how your faculty plans, drafts, checks and decides, with a new assistant in the loop. The frameworks, the data line, the disclosure norms and the PD are the durable part. The tool is replaceable; the practice and the accountability are not.

   Note: If you have any stake in a tool you are recommending — even an indirect one — that is a conflict of interest to manage openly. We deal with how to document and disclose that in Module 4. Flag it to yourself now.

   Activity — Write your one-paragraph mandate (10 min)

   In your own words, write a single paragraph you could say out loud to your faculty that answers three questions: (1) Why are we addressing AI now (not later, not never)? (2) What two harms are we steering between? (3) What is my role as the leader of this? Keep it to ~120 words — short enough to actually say in a meeting. This paragraph becomes the opening of your position statement in the capstone.

   Knowledge check

   1The Australian Framework puts accountability with 'teachers and leaders'. Practically, what does that mean a leader cannot do?

   Show answer

   2Why is 'doing nothing' a risk rather than a safe, neutral default?

   Show answer

   3A colleague says 'let's just buy the best AI tool and get everyone on it'. Why is that the wrong frame for what you are leading?

   Show answer

   Mark this module complete
2. 2

   Setting faculty & school AI policy that people will actually follow

   Align to the Australian Framework's six principles and NSW DoE guidance, make the six ethical checks a staff norm, set disclosure expectations, hold the data line — and write it on one page.~60 min

   By the end of this module you'll be able to:

   • Draft a faculty AI position aligned to the Australian Framework's six principles and NSW DoE guidance.
   • Translate the DoE's six ethical checks into a staff norm and set a clear disclosure expectation.
   • Produce a position that fits on one page — short enough that staff will actually read and follow it.

   Standards7.2 Comply with legislative, administrative and organisational requirements7.1 Meet professional ethics and responsibilities4.5 Use ICT safely, responsibly and ethically

   Short beats comprehensive

   The most common policy mistake leaders make is writing a thorough document nobody reads. A 30-page AI policy that sits in a drive is worth less than a one-page position that staff can recall in a meeting. Your aim is a usable, living position — aligned to the frameworks above you, written for the people below you. If it does not fit on a page, it will not change behaviour.

   You are not writing this from scratch. Two documents do the heavy lifting; your job is to localise them to your faculty.

   Anchor 1 — the Australian Framework's six principles

   Every line of your position should trace back to one of these. Keep the principle names visible so staff see the why:

   • Teaching & Learning — AI enhances teacher expertise and builds students' AI literacy, creativity and critical thinking.
   • Human & Social Wellbeing — it must not harm anyone's safety, dignity or wellbeing.
   • Transparency — the school community knows how and when AI is used.
   • Fairness — accessible, inclusive, free from bias, and respectful of Indigenous Cultural and Intellectual Property (ICIP).
   • Accountability — teachers and leaders remain responsible for the decisions AI supports.
   • Privacy, Security & Safety — student data is protected.

   Anchor 2 — NSW DoE guidance and the six ethical checks

   DoE recommends its own secured tool, NSWEduChat (available to all staff and Year 5–12 students), and sets minimum safety practices: strip sensitive or identifying information out of general/public tools (NSWEduChat permits limited personal information because it is secured), review and edit every output, and check for errors, bias and harmful content. Make the DoE's six ethical checks an explicit staff norm — not a poster, a habit:

   Check	The question a teacher asks before trusting an output
   Oversight	Did I actively review this against what I intended?
   Diversity	Did I seek varied perspectives, not one narrow take?
   Explainability	Can I explain how I produced this?
   Knowledge boundaries	Am I only using AI where I can judge whether it is right?
   Respect for others	Have I protected others' data and dignity?
   Community alignment	Does this fit our school community's values?

   The hard line — and why it is non-negotiable

   Your position must state the data line in words a tired teacher cannot misread:

   Never enter personal or identifying information about a student into a general AI tool. No names, NESA/student numbers, dates of birth, addresses, health, disability, behaviour or wellbeing records.

   This is the Privacy and Personal Information Protection Act 1998 (NSW) and the Child Safe Standards, not etiquette. The single nuance: NSWEduChat is secured, so DoE guidance permits limited personal information there; in any public tool, strip it out. Lessio removes the dilemma — it is built student-data-free, so staff describe a cohort ("mixed-ability Year 9, six EAL/D learners"), never a child. As a leader, hold this line without exception — one breach is a reportable matter, not a teachable moment.

   Set the disclosure expectation

   Transparency is a principle, so make it operational. Decide and write down: when staff use AI to produce something that reaches students or families (a report comment bank, a unit, a newsletter), what is the disclosure norm? A workable default: AI may assist drafting; a teacher reviews, edits and owns the final product, and we are open about using it. For student assessment, disclosure is NESA's territory and lives in your assessment guidance — covered in Modules 4 and 5 — but the staff-facing norm starts here.

   A one-page skeleton you can fill

   Your position statement, on a single page, needs only these headings:

   1. Why — your one-paragraph mandate from Module 1.
   2. Principles we work under — name the Australian Framework's six.
   3. Approved tools — e.g. NSWEduChat for general tasks; Lessio for planning; "ask before using anything else".
   4. The data line — the hard rule, verbatim, with the NSWEduChat nuance.
   5. How we check outputs — the six ethical checks as our habit; a human signs off.
   6. Disclosure — our norm for being open about AI use.
   7. Who to ask — the named person (you) for questions and exceptions.

   Keep it a living position: version it, date it, and review it (Module 5). "Draft v1, reviewed each semester" beats a polished document that ossifies.

   Activity — Draft two clauses of your position (12 min)

   Write two sections of your one-page position in full sentences your staff would actually read: (1) the Approved tools clause (which tools for which jobs, and what to do about anything else), and (2) the data line clause (verbatim hard rule + the NSWEduChat nuance). Hold them to a few sentences each. These plug straight into your capstone.

   Knowledge check

   1Why is a one-page position usually more effective than a comprehensive AI policy document?

   Show answer

   2Your position says 'no identifying student information in AI tools', but a head teacher points out a colleague uses NSWEduChat. How do you reconcile that?

   Show answer

   3How do the DoE's six ethical checks function in a faculty position — as decoration or as something more?

   Show answer

   Mark this module complete
3. 3

   Bringing staff with you — change leadership, not a memo

   The predictable camps, leading by modelling, using the flagship PD as the vehicle, building psychological safety, and addressing the deskilling fear honestly.~55 min

   By the end of this module you'll be able to:

   • Map your staff into readiness groups and plan a differentiated response to each.
   • Plan to lead the change by modelling and by deploying structured PD, not by mandate.
   • Address the deskilling fear honestly and build the psychological safety the change needs.

   Standards6.3 Engage with colleagues and improve practice6.2 Engage in professional learning6.4 Apply professional learning and improve student learning

   A policy does not change practice — people do

   You can write the perfect one-page position and still change nothing. Adoption is a change-leadership task, and people do not adopt change because they were sent a memo. They adopt it when they understand the why, feel safe to try, see their leader doing it, and are supported when they struggle. Your position from Module 2 is the what; this module is the how you bring people to it.

   Know your camps — then differentiate

   Any staffroom sorts itself into predictable groups when you introduce AI. Name them so you can lead each one differently rather than pitching to an imaginary average:

   Camp	Stance	What they need from you
   Eager / over-trusting	"This is brilliant — I use it for everything."	Guard-rails, not encouragement. Channel their energy; make them model careful use, not just enthusiastic use. Watch for the data line slipping.
   Curious / cautious	"I'm interested but unsure if I'm allowed / doing it right."	The biggest group, and your priority. Clear permission, a safe first task, and structured PD. Most of your effort goes here.
   Resistant / fearful	"It's cheating / a fad / it'll deskill us / replace us."	To be heard, not overridden. Honest answers to real fears, and no mandate. Some of their caution is wisdom you want.

   Resistance is data, not defiance. The resistant colleague who says "this will deskill our new teachers" has named a real risk you should be managing — thank them and use it.

   Lead by modelling

   Staff watch what leaders do, not what they circulate. If you want careful, transparent, teacher-in-the-loop use, be seen doing exactly that: show a unit you drafted with Lessio and the changes you made to it; narrate the ethical checks out loud; say openly when AI got something wrong and you fixed it. Modelling careful use — including the verifying and the editing — is the single most persuasive move you have. The opposite is also true: if you mandate AI but never touch it, staff read the signal accurately.

   Use structured PD as the vehicle

   Do not ask staff to "go figure out AI". Give them a credible, NSW-grounded path. The flagship course, "Teaching with AI: Ethical & Effective Practice", is built to be that vehicle — and because it is Standards-relevant PD, the time counts toward each teacher's 100 NESA maintenance hours, self-logged in eTAMS. Running structured PD does three things at once: it builds capability, it signals that the school takes this seriously enough to invest time in, and it gives the cautious majority a safe, shared starting point. (Planning the day itself is the next section.)

   Build psychological safety

   People will not experiment in front of colleagues if they fear looking incompetent or being caught out. Your job is to make trying-and-refining safe:

   • Normalise not-knowing. "None of us were trained for this" is true and disarming.
   • Separate learning from appraisal. Make it explicit that AI PD is developmental, not something that feeds performance management.
   • Reward disclosure, not concealment. A teacher who says "I used AI and it gave me a wrong worked example, so I fixed it" is doing exactly what you want — celebrate that, do not police it.

   Address the deskilling fear honestly

   This is the most legitimate objection, and waving it away costs you credibility. Be straight: over-reliance can erode skill and judgement, especially for early-career teachers. That is precisely why the model is teacher-in-the-loop — AI drafts, a professional decides — and why you verify rather than assume. Framed honestly, the deskilling fear becomes the argument for your guard-rails, not against adoption. Tell staff plainly: we use AI to remove drudgery so you can spend judgement where it matters, not to replace the judgement itself.

   Activity — Map your staff and write one message (12 min)

   (1) Sort your actual faculty into the three camps — roughly how many in each? (2) Pick the cautious majority and draft the opening of a short staff message that gives them permission, names the safe first task, and points to the PD. Three or four sentences, warm and concrete. This becomes part of your rollout plan in the capstone.

   Knowledge check

   1Why might a technically perfect AI policy still fail to change how your faculty works?

   Show answer

   2A respected senior teacher resists, saying AI will deskill your early-career staff. What is the leaderly response?

   Show answer

   3Why is leading by modelling more powerful than mandating AI use?

   Show answer

   Mark this module complete
4. 4

   Privacy, safety & compliance at the school level

   Your legal obligations as a leader — PIP Act, Child Safe Standards, Disability Standards — tool and vendor approval, data hygiene, governance, and managing conflicts of interest.~55 min

   By the end of this module you'll be able to:

   • Identify the legislation and standards that bind AI use in your school and what each requires of you.
   • Apply a tool/vendor approval discipline, distinguishing a secured tool from a general one.
   • Document a decision — including any conflict of interest — to a standard that survives scrutiny.

   Standards7.2 Comply with legislative, administrative and organisational requirements4.5 Use ICT safely, responsibly and ethically7.1 Meet professional ethics and responsibilities

   As a leader, the compliance burden is yours to hold

   A classroom teacher applies the rules; a leader is accountable for the system that makes the rules followable. That means knowing the legal frame, deciding which tools are allowed, setting data hygiene, and being able to show — later, to a parent, your principal, or a regulator — that decisions were made carefully. This module is the unglamorous, load-bearing part of leading AI well.

   The legal frame, in plain terms

   Three instruments sit behind your AI position. You do not need to be a lawyer, but you must know what each demands:

   Instrument	What it requires of you
   Privacy & Personal Information Protection Act 1998 (NSW)	Protect students' personal information. In AI terms: identifying student data must not go into general tools. A breach is a reportable matter, not an internal hiccup.
   Child Safe Standards	Children's safety and wellbeing come first. AI use must not expose students to harm, unsafe content, or misuse of their information.
   Disability Standards for Education 2005	Reasonable adjustments are a legal requirement, not a kindness. AI must support inclusion and accessibility — and must never become a barrier or a substitute for an adjustment a student is entitled to.

   The Disability Standards cut both ways: AI can help you produce adjusted materials faster, and you must ensure AI-mediated tasks remain accessible. Lead it as an inclusion opportunity with an inclusion obligation attached.

   Tool and vendor approval — secured vs general

   Not all tools are equal, and "the staff can use whatever they like" is not a defensible position. Set an approval discipline:

   • NSWEduChat is the DoE's department-built, secured tool — available to all staff and Year 5–12 students. Because it is secured, DoE guidance permits limited personal information. It is your safe default for general tasks.
   • General/public tools are unsecured. Identifying student data must never enter them. If staff want a new tool, it goes through you — not onto the staffroom grapevine.
   • Lessio is student-data-free by design: it drafts to the verbatim NESA syllabus and NSW DoE standards from a cohort description, never a child's record — which removes the data question from planning entirely.

   Make the rule simple: approved tools are named in the position; anything else requires a conversation before use. That single sentence converts shadow procurement into a governed decision.

   Data hygiene as a standing habit

   Set expectations that outlast any one tool: strip identifying information before using a general tool; prefer secured or student-data-free tools for anything touching real cohorts; review and edit every output before it reaches a student; and never paste a whole class list, a report draft, or a wellbeing note into a chatbot. Make these the faculty's reflexes, audited occasionally, not one-off instructions.

   Governance, conflicts of interest and documenting decisions

   This is where leaders are most exposed and least prepared. When you adopt or recommend a tool, govern the decision and write it down.

   • Document the decision. What did you approve, when, on what basis, against which principles? A short, dated record is your defence if the decision is questioned later. Memory is not a record.
   • Disclose conflicts of interest. If you — or a relative, or a colleague making the call — have any stake in a tool being adopted (financial, developmental, reputational, or a side venture), that is a conflict of interest. It does not necessarily bar the tool, but it must be disclosed in writing and the decision kept at arm's length — ideally made by someone without the interest, against objective criteria. Handled openly, a disclosed interest is manageable; concealed, it is a probity failure.
   • Keep procurement honest. Evaluate tools against your principles and the legal frame, not against who is enthusiastic or who knows the vendor.

   The test to apply to any adoption decision: if this were read out at a P&C meeting or in an audit, would it stand up? If yes, you have governed it well. If you would wince, document and disclose more.

   Activity — Draft your approval rule and a decision record (12 min)

   (1) Write the one-sentence tool-approval rule for your position (approved tools named; anything else needs a conversation). (2) Draft a short decision record template — the four or five fields you would log for any AI tool you approve (what, when, basis, principles met, any conflict of interest disclosed). You will reuse the template every time you adopt something.

   Knowledge check

   1Beyond privacy, the Disability Standards for Education 2005 are relevant to AI adoption. How — and why does it 'cut both ways'?

   Show answer

   2A leader wants to adopt an AI tool in which they hold a personal stake. What does responsible governance require — and does the interest disqualify the tool?

   Show answer

   3Why insist on a short, dated written record for every AI tool you approve?

   Show answer

   Mark this module complete
5. 5

   Measuring impact & making it stick

   Decide what to measure (and what not to do), protect the reclaimed time, and embed AI in your professional-learning plan and eTAMS so it outlasts the initial enthusiasm.~50 min

   By the end of this module you'll be able to:

   • Choose meaningful measures of impact — workload, planning quality, staff wellbeing — and reject vanity metrics.
   • Protect reclaimed time and embed AI in the school's professional-learning plan and eTAMS.
   • Establish a review cycle that iterates the position rather than letting it ossify.

   Standards3.6 Evaluate and improve teaching programs6.4 Apply professional learning and improve student learning6.2 Engage in professional learning

   What gets measured signals what you value

   A rollout with no measurement drifts: enthusiasm fades, practice quietly reverts, and you cannot tell your principal whether any of it worked. But measuring the wrong things is worse than measuring nothing — it pushes staff toward gaming a number. As the leader, choose a small set of measures that reflect why you did this in the first place.

   Measure what matters — not what is easy to count

   Tie measures to your purpose: less drudgery, better teaching, healthier staff.

   Worth measuring	Why	A simple way to read it
   Workload / hours saved	The core promise — admin time returned to teachers.	Brief before/after estimates on specific tasks (e.g. units, comment banks); ask staff, periodically.
   Quality of planning & resources	Time saved is hollow if quality drops.	Sample programs against your template/standards; faculty moderation.
   Staff wellbeing & confidence	Adoption should reduce pressure, not add it.	Short pulse check: do staff feel more supported, less overloaded, more confident?

   Avoid the vanity metric of "how many staff are using AI". Usage is an input, not an outcome — high usage with poor judgement is a problem, not a success. Measure whether the work and the people are better off.

   The trap that quietly kills the benefit

   Name this one explicitly, because it is the most common way AI initiatives fail their staff:

   If you let the time AI saves simply refill with more work, you have not helped anyone — you have just raised the baseline.

   Reclaimed time is the return on this change. If a teacher saves three hours on programming and you hand back three hours of new tasks, the wellbeing benefit evaporates and the staff conclude — correctly — that "efficiency" meant "do more". As the leader, protect the reclaimed time on purpose. Be explicit that hours saved are meant to ease load, deepen planning, or go to students — not to absorb more administration. This is a leadership choice, not an automatic outcome.

   Make it stick — embed, do not bolt on

   Initiatives that live in one keen leader's head die when that leader moves rooms. Embed AI into the structures that persist:

   • The school's professional-learning plan. Put AI capability in the plan as ongoing PD, not a one-off SDD. Because it is Standards-relevant, staff log the hours in eTAMS toward their 100-hour NESA maintenance requirement — so it counts, formally, with no endorsement gate (the Accredited/Elective categories were removed in August 2024).
   • Existing routines. Fold AI practice into faculty meetings, programming days and induction for new staff, so it becomes "how we work", not "the AI thing we did".
   • A named owner. Someone — often you — keeps a light hand on it: answering questions, watching the data line, refreshing the PD.

   Review and iterate the position

   Your one-page position from Module 2 is a draft v1, not tablets of stone. Set a review rhythm — each semester is sensible — and at each review ask: Has the tooling changed? Has DoE or NESA guidance shifted? What is the data telling us? Where are staff still unsure? Then update, re-date, and re-communicate. A position that visibly evolves keeps its authority; one that ossifies gets ignored.

   Activity — Choose three measures and write your protection statement (12 min)

   (1) Pick three measures you will actually track (one workload, one quality, one wellbeing) and, for each, one realistic way to gather it. (2) Write one sentence — your time-protection statement — that tells staff what the hours AI saves are for, and what they will not be refilled with. Both go into your rollout plan in the capstone.

   Knowledge check

   1Why is 'percentage of staff using AI' a poor measure of whether your rollout is working?

   Show answer

   2AI saves your teachers several hours a week. What is the leadership trap, and what must you do about it?

   Show answer

   3How does embedding AI PD in the professional-learning plan and eTAMS help it 'stick', and why does it count?

   Show answer

   Mark this module complete
6. 6

   Capstone — your faculty AI position & rollout plan

   Bring it together: draft a one-page faculty AI position statement and a staff rollout plan — including a staff development day outline and how you will measure success — then self-assess and log it.~60 min

   By the end of this module you'll be able to:

   • Produce a one-page faculty/school AI position statement aligned to the Australian Framework and DoE guidance.
   • Produce a staff rollout plan including a staff development day outline and success measures.
   • Self-assess your plan against the leader's checklist and log the work as Lead-level PD in eTAMS.

   Standards7.1 Meet professional ethics and responsibilities6.3 Engage with colleagues and improve practice3.6 Evaluate and improve teaching programs

   From learning to leadership artefact

   Across five modules you have drafted the pieces: a mandate, policy clauses, a staff message, an approval rule and decision record, measures and a time-protection statement. The capstone assembles them into two artefacts you can take to your executive and actually run. This is the point of the course — not knowledge for its own sake, but a plan you can lead from on Monday.

   Part A — Your one-page faculty AI position statement

   Assemble the one-page skeleton from Module 2 into a finished draft. On a single page:

   1. Why — your one-paragraph mandate (Module 1).
   2. Principles we work under — the Australian Framework's six, named.
   3. Approved tools — e.g. NSWEduChat (general), Lessio (planning); anything else needs a conversation.
   4. The data line — the hard rule verbatim, with the NSWEduChat nuance.
   5. How we check outputs — the DoE's six ethical checks as our habit; a human signs off on everything.
   6. Disclosure — your norm for being open about AI use.
   7. Review — version, date, and the cycle (e.g. each semester).

   The one-page test: if it spills past a page, cut — do not append. The discipline is the deliverable.

   Part B — Your staff rollout plan

   A short plan (a page or two) for bringing people to the position. Cover:

   • Staff map. Your three camps (eager / cautious / resistant), roughly sized, with how you will respond to each — most effort on the cautious majority.
   • The message. Your opening staff communication: permission, a safe first task, the why, and the PD path.
   • The PD vehicle. How staff build capability — the flagship "Teaching with AI" course as the spine — logged in eTAMS toward the 100 NESA maintenance hours.
   • An SDD outline. A run-sheet for a staff development day (see the skeleton below).
   • Compliance & governance. Your tool-approval rule and decision-record template (Module 4); how the data line is held.
   • Measures. Your three measures (workload / quality / wellbeing) and your time-protection statement.
   • Review. When and how you will revisit the position (Module 5).

   A staff development day (SDD) outline you can adapt

   A workable half-day shape — adjust to your context:

   Segment	Focus
   Open (15 min)	Your mandate — why now, the twin failure modes, your role. Set a non-judgemental tone.
   The frameworks (30 min)	The Australian Framework's six principles, the DoE's six ethical checks, and the data line.
   Model it (20 min)	Show a real AI-drafted artefact and the human edits you made — narrate the verifying.
   Hands-on (45 min)	Staff try a safe, real, de-identified task (e.g. draft a unit in Lessio, reword in NSWEduChat) and debrief 'kept / changed'.
   Our position (20 min)	Walk the one-page position; agree the disclosure norm; confirm approved tools.
   Close & log (15 min)	Each teacher writes a reflection for their eTAMS record; you name the next step and the named owner.

   Self-assessment — the leader's checklist

   Before you call this done, score your two artefacts against the leader's responsible-adoption checklist (it is in the course Checklist). If any item is missing, that is your next edit, not a footnote.

   Log it as Lead-level PD

   This work is Standards-relevant and squarely at the Lead career stage — you are establishing policy, leading colleagues, and evaluating practice (APST Standards 6 and 7, and 3.6). Log your hours and a short reflection in eTAMS. No endorsement gate applies; it counts toward your 100 maintenance hours like any Standards-relevant PD.

   Activity — Finish and pressure-test your plan (20 min)

   Complete Part A and Part B into clean drafts. Then pressure-test them against two questions: (1) Could a relief head teacher pick up this position and apply it tomorrow? (2) If my decisions here were read aloud at a P&C meeting or an audit, would they stand up? Fix anything that fails either test. Save both artefacts and log the PD in eTAMS.

   Knowledge check

   1What are the two artefacts the capstone asks you to produce, and who is each one for?

   Show answer

   2Why does the capstone insist the position pass a 'relief head teacher could apply this tomorrow' test?

   Show answer

   3At which APST career stage does this capstone sit, and why does it count as PD without an endorsement gate?

   Show answer

   Mark this module complete

Take-away prompt library

Ready, RICE-shaped prompts for common NSW teaching jobs (Module 3). De-identified — copy one, swap in your details, and use it today.

Role: You are an experienced NSW secondary head teacher writing internal policy. Input: Our faculty is [subject], [number] staff, mostly [readiness — e.g. cautious]. We work under the Australian Framework for Generative AI in Schools (six principles) and NSW DoE guidance; approved tools are NSWEduChat (general tasks) and Lessio (planning); our hard rule is no identifying student information in general tools. Context: This must fit on ONE page and be read by busy teachers. Expectations: Draft a one-page position with these headings — Why; Principles we work under (name the six); Approved tools; The data line (state it as an unmissable rule); How we check outputs (the DoE's six ethical checks as a habit); Disclosure; Who to ask. Plain English, no legalese. Self-check: After drafting, list any sentence a tired relief teacher could misread, and tell me where I must insert my own school's specifics. Remind me this is a draft to edit, shorten and own — not to publish as-is.

Role: You are a NSW school leader writing a warm, credible staff email. Input: We are introducing a considered approach to AI — not a mandate. Key points: AI is for cutting admin so staff can spend judgement where it matters; teacher-in-the-loop (AI drafts, a professional decides); the student-data hard line; structured PD is provided and counts toward NESA maintenance hours in eTAMS; I will model careful use myself. Context: Audience is the cautious majority — interested but unsure if they are 'allowed' or doing it right; tone is reassuring, not hype, and explicitly safe to experiment. Expectations: Draft a short email (~200 words) giving permission, naming one safe first task, the why, and the PD path, signed off with who to ask. Self-check: Flag any sentence that sounds like a mandate or like pressure, and any claim I should soften or verify against our actual context before sending.

Role: You are a NSW head teacher planning a staff development day. Input: Half-day session to launch our faculty AI approach. Must cover: my leadership mandate (why now, the twin failure modes); the Australian Framework's six principles; the DoE's six ethical checks; the student-data hard line; live modelling of careful, edited AI use; a hands-on activity on a safe, de-identified real task (drafting a unit in Lessio, rewording in NSWEduChat); walking our one-page position; and a close where staff log a reflection for eTAMS. Context: Mixed-readiness staff; tone non-judgemental and practical; not a tech-bashing or a hype session. Expectations: Produce a timed run-sheet (segment, minutes, purpose, facilitator note) totalling ~3 hours, plus a short list of what I must prepare beforehand (logins, a real artefact to model, the position projected). Self-check: Identify the segment most likely to run over or get resistance, and suggest how I would handle it; flag anything I should adapt to my own faculty.

Role: You are a NSW school leader writing a short, plain-English explainer for parents and carers. Input: We use AI to help staff reduce administrative work (drafting and planning), always with a teacher reviewing and owning the final product. We protect student information — identifying details are never put into general AI tools. We follow the national Australian Framework for Generative AI in Schools and NSW Department of Education guidance, and our staff complete professional learning on using it responsibly. For student assessment, our school decides where AI is and is not permitted, and we uphold academic integrity. Context: Audience is parents with no technical background; aim is reassurance and transparency, not jargon. Expectations: Draft ~250 words covering how the school uses AI, how student data is protected, who remains accountable (teachers and leaders), and where to direct questions. Calm and non-hype. Self-check: List any sentence that over-promises or could be misread as 'AI teaches the students', and tell me which details I must localise (contact point, assessment specifics) before publishing.

Role: You are advising a NSW school executive on governance. Input: We are adopting AI in a faculty under the Australian Framework, NSW DoE guidance, the Privacy and Personal Information Protection Act 1998 (NSW), the Child Safe Standards and the Disability Standards for Education 2005. Context: We need a starting risk register to manage, not a comprehensive audit. Expectations: Produce a table of the most important risks (e.g. student data entering general tools; hallucinated/biased content reaching students; over-reliance and deskilling; assessment integrity; inaccessible AI-mediated tasks under the Disability Standards; conflict of interest in tool adoption; reclaimed time being refilled). For each: a plain description, likely impact (low/med/high), and a concrete mitigation a head teacher can actually implement. Self-check: Flag any risk where the mitigation depends on a decision only my principal or the DoE can make, and note any risk specific to my context that I should add myself.

Role: You are a NSW school leader writing a professional-learning-plan entry. Input: Staff are completing structured AI professional learning (the 'Teaching with AI' course and this leadership course) and applying it to faculty practice. Context: It must be embedded as ongoing PD (not a one-off), be Standards-relevant, and be self-logged in eTAMS toward the 100-hour NESA maintenance requirement — no endorsement gate since the August 2024 change. Expectations: Draft a plan entry stating the goal, the activities, the APST descriptors addressed (favouring 6.2, 6.3, 6.4, 7.1, 7.2, 4.5, 3.6 — pick those that genuinely fit), how it embeds in existing routines (meetings, programming days, induction), the success measures (workload / planning quality / wellbeing — not raw usage), and the review cycle. Self-check: For each Standard I have claimed, give one sentence justifying the link; flag any descriptor that looks like a stretch so I can drop it, and remind me to localise the measures to my faculty.

Standards-relevant professional learning, mapped to the APST · content verified against national and NSW frameworks, June 2026 · self-log the hours in eTAMS.