For Australian principals, school executives, curriculum leaders and governing bodies. Updated 26 October 2026. This is a planning checklist, not a statement that one AI policy applies identically to every state, territory or school sector. Confirm current local requirements before implementation.
Artificial intelligence is already changing how Australian students research, draft, revise and ask for help. Yet purchasing a chatbot, issuing a staff memo or banning a particular tool does not make a school AI-ready. The real test is whether school leaders can protect student wellbeing and privacy, safeguard the integrity of assessment, support teachers and demonstrate that students are still doing meaningful intellectual work.
To get Australian schools AI-ready for 2027, leaders need a documented decision-making process, approved uses and tools, redesigned assessment, staff capability, family communication and evidence of student learning. This guide translates those priorities into a usable leadership checklist and a 90-day preparation plan.
The starting point is Australia's national Framework for Generative AI in Schools. It sets out six guiding principles: teaching and learning; human and social wellbeing; transparency; fairness; accountability; and privacy, security and safety. Education Ministers endorsed its 2024 review in June 2025. The framework is national guidance; implementation also depends on the rules of each jurisdiction and school sector.
AI readiness in 2027 is a leadership question, not a software question
Consider a secondary school with five departments. English teachers worry that students will outsource essays. Science teachers want AI-generated feedback on explanations. Learning support staff are interested in accessibility tools. Administration teams want help drafting newsletters. Families are unsure which services children may use at home. Everybody is responding to the same technology, but not to the same educational risk.
A single list of approved brands cannot resolve these differences. A system suitable for generating fictional examples may be unsuitable for processing identifiable assessment work. An AI tool that speeds up planning may produce errors, stereotypes or inaccessible language. A polished piece of student work may conceal weak understanding.
That is why the first question should not be Which AI platform should we buy? It should be Which human capabilities are we trying to strengthen, and how will we know?
The Australian Education Research Organisation (AERO) emphasises protecting evidence-based teaching, identifying risks to learning, setting clear expectations, reconsidering assessment and applying a culturally responsive lens. It also cautions that evidence about AI's impact on school learning outcomes remains limited. Enthusiasm and anecdote cannot substitute for evaluation.
The 12-point AI-ready school leadership checklist
Use the following table in a leadership meeting. For each item, record an accountable owner, a concrete piece of evidence and a date for review. A tick is meaningful only if it can be substantiated.
| Leadership requirement | Evidence to have before rollout | Suggested owner |
|---|---|---|
| 1. Define educational purpose | Two or three explicit learning problems, not a product wish list | Principal and learning leaders |
| 2. Confirm applicable rules | Current sector and jurisdiction guidance recorded | Principal or compliance lead |
| 3. Establish governance | Named decision makers, escalation paths and review dates | Executive team |
| 4. Map AI use cases | Approved, conditional and prohibited uses by task | Curriculum leaders |
| 5. Vet products and data | Documented privacy, security, age and accessibility review | ICT, privacy and procurement |
| 6. Redesign assessment | Task-level AI conditions and authorship evidence | Heads of learning areas |
| 7. Prepare staff | Practice-based professional learning and support | Professional learning lead |
| 8. Teach student AI literacy | Age-appropriate lessons and examples of permitted use | Teachers and wellbeing teams |
| 9. Protect inclusion and wellbeing | Accessible alternatives and safeguards for vulnerable students | Inclusion and wellbeing leaders |
| 10. Inform families | Plain-language guidance and a clear contact route | Principal and communications |
| 11. Pilot for learning | Baseline, comparison and agreed measures | Teaching and learning team |
| 12. Review and improve | Incident log, feedback, decisions and scheduled policy updates | Governance group |
The checklist is not an official government compliance instrument. It is an operational planning tool aligned with the national framework and established school leadership responsibilities.
1. Start with the six national principles—and translate each into a decision
School leaders should give each principle an operational question rather than treating it as a wall poster.
Teaching and learning: Does the proposed use help learners practise, explain, transfer or retain knowledge? For example, an AI tutor might ask a student to explain a fraction method before offering feedback. If the tool simply supplies the worked solution, the learning purpose may be undermined.
Human and social wellbeing: What might the use do to student confidence, relationships, workload and safety? A support chatbot is not a substitute for a trusted adult or a school's safeguarding processes.
Transparency: Can students, families and staff explain when AI is used, why, and what remains human work? A student should not need to guess whether AI-written text is permitted in an assignment.
Fairness: Could the use systematically disadvantage students with disabilities, multilingual learners, students without home devices or particular cultural communities? Accessibility and equity require testing, not just good intentions.
Accountability: Who approves a use case, checks output, responds to errors and decides to discontinue it? Responsibility remains with appropriately authorised people, not an algorithm.
Privacy, security and safety: What personal information enters the system, where it goes, who can access it and what happens if the product changes? Never assume that a convenient classroom tool has been approved to process identifiable student information.
These questions keep a school aligned with the framework while allowing sensible local decisions. Leaders should consult their own sector's current policies before authorising classroom or administrative deployment.
2. Build a school AI register before expanding use
A basic register is more valuable than a broad promise to “use AI responsibly”. It gives staff a transparent answer to the question: What is allowed here?
Record each proposed use with the following fields:
The learning or administrative purpose and why AI is needed.
The approved product, age group, user role and accessibility requirements.
The data that may and may not be entered.
The teacher's or leader's review responsibilities.
What students must disclose when AI contributes.
How the school will evaluate benefits and harms.
Who can suspend the use and when it will be reviewed.
For example, generating teacher-reviewed sample questions from public curriculum descriptions is a different use case from automatically evaluating named students' assessments. Both may involve text generation, but their consequences and safeguards differ substantially.
The Safer Technologies 4 Schools framework is one source for structured assessment of education technologies. Its July 2026 update includes strengthened AI-related safety considerations. A school's product review should draw on available sector processes rather than treating a vendor's marketing claims as independent verification.
Also conduct a short “shadow AI” check: ask staff anonymously which services they already use for preparation, feedback and administration. The aim is to identify training and governance gaps, not to embarrass teachers. Unexpected features embedded in existing platforms should be included.
3. Redesign assessment to reveal actual learning
Generative AI makes the distinction between a convincing product and student understanding especially important. An elegant submission is not reliable evidence that the learner can explain its argument, solve a related problem or defend its sources.
For 2027, audit the highest-risk assessments first. Ask: What knowledge or skill is this task intended to assess? Which parts could be delegated to AI without demonstrating that skill? How will the teacher authenticate the student's work fairly?
Practical approaches include drafts produced or discussed in class, short oral explanations, annotated decisions, process journals, retrieval checks and new problems that require applying what was learned. None is automatically foolproof. Multiple complementary pieces of evidence are stronger than one supposedly “AI-proof” assignment.
There is a recent and specific Australian example. On 1 September 2026, the NSW Government announced changes to take-home HSC assessment, limiting the number and weighting of such tasks for relevant cohorts while retaining assessment options such as in-class tasks, oral components and process diaries. Certain major works and subjects have exceptions or additional authentication requirements. This is a NSW HSC measure, not a nationwide rule for all Australian schools and year levels.
Do not build disciplinary decisions on an AI-detector score alone. A school should follow its applicable assessment, procedural fairness and academic integrity requirements, investigate evidence proportionately and provide students an opportunity to respond.
4. Teach AI literacy as thinking, not prompt tricks
Students deserve more than a poster warning them not to copy answers. They need explicit instruction about AI limitations, fabrication, bias, attribution, privacy and when choosing not to use AI is the better learning decision.
A useful progression can begin with teacher-led demonstrations of a plausible but incorrect response. Students compare the output with a reliable source, mark unsupported claims and explain the correction. Older students can investigate which tasks benefit from feedback and which require independent practice before assistance.
Try a simple three-stage classroom routine:
Before AI: The learner states their initial answer, working or question in their own words. The teacher observes the starting point.
With approved AI assistance: The learner requests a hint, counterexample or critique—not a replacement submission. They verify important claims independently.
After AI: The learner closes the tool and explains or applies the idea without its help. The teacher checks what can now be done independently.
This sequence is not a mandated Australian curriculum model. It is an illustrative learning-centred practice. The Australian Curriculum's AI connection offers context for understanding AI and its limitations, while schools must select age-appropriate teaching strategies consistent with local requirements.
5. Give teachers practical professional learning, not a one-off demonstration
A staff session showing twenty prompts may be enjoyable but leave teachers without a repeatable decision process. Professional learning is stronger when it deals with the real tasks educators face: a suspicious assessment, a pupil who trusts an incorrect answer, an accessibility request or an inappropriate data upload.
Structure teacher learning around three practices. First, participants examine sample AI outputs for accuracy, bias and curriculum fit. Second, teams redesign one lesson or assessment with clear boundaries and a learning check. Third, colleagues rehearse a short conversation with a student about authorship and responsible assistance.
Teachers should have room to say that an intended AI use provides no clear benefit. They should also receive consistent advice on approved services, disclosure, privacy incidents, and escalation. Accountability without professional support creates hidden workarounds.
Where teachers are experimenting, build peer observation and reflection into the process: What changed in students' questions, explanations, independence or misconceptions? What was easier for the teacher, and what new checking work appeared? Decisions should be made from that evidence, not from the length of time students spend in a platform.
6. Put inclusion, privacy and wellbeing ahead of rapid rollout
A school can unintentionally widen existing gaps if its AI plan assumes reliable internet, paid accounts, fluent English or unrestricted access to the latest devices. Maintain a meaningful non-AI route to learning and assessment wherever necessary. Consultation with disability and inclusion specialists matters before a tool becomes a default.
Cultural responsiveness must also be an explicit part of the review. For Aboriginal and Torres Strait Islander learners and communities, consider whether AI-generated material misrepresents histories, languages, knowledge systems or community perspectives. Teacher oversight and appropriate community consultation are especially important when content claims to represent cultural knowledge.
Privacy reviews should examine terms, age eligibility, consent obligations, retention, training on uploaded material, contractual arrangements and breach response. If a tool has not been approved for identifiable student data, staff should not paste student reports, photographs, sensitive wellbeing notes or marked work into it.
The NSW Department of Education's public guidance illustrates the importance of approved tools and safe handling of information. Victorian government guidance on appropriate GenAI use also highlights disclosure, accountability and monitoring. The details differ by jurisdiction: check the rules that govern your school.
7. Communicate one clear set of expectations to families
Parents often encounter AI through homework before they encounter the school's policy. Send a short, readable guide addressing which tools are approved, what constitutes acceptable help, how students should acknowledge assistance, and whom to contact if a problem arises.
Replace abstract warnings with examples. “Ask an approved tool to quiz you on your notes, then answer independently” is more helpful than “use AI ethically”. Conversely, “do not submit an AI-generated response as if you wrote it” is clearer than a vague demand for originality.
Invite families to raise accessibility, language, device access and safety concerns. Do not shift the full burden of monitoring AI use onto parents. Assessment rules should be realistic in students' actual circumstances and consistently reinforced at school.
A 90-day preparation plan for Term 1, 2027
Schools with limited capacity do not need to solve every future question at once. They need a controlled sequence of decisions.
Days 1–30: Discover and set boundaries
Form a small representative group including teaching, ICT, wellbeing, inclusion, administration and leadership. Review national guidance alongside current sector rules. List existing AI use, staff needs and assessment vulnerabilities. Identify a narrow set of potential benefits. Establish prohibited data practices and an escalation process immediately.
Deliverables: governance owner, initial use-case register, privacy questions, draft student/staff expectations and an assessment risk list.
Days 31–60: Train and pilot cautiously
Select one or two low-risk, clearly defined use cases. Agree which teachers and students may participate, which alternatives remain available, and what consent or other approvals may be required. Run short professional learning sessions. Document baseline performance before introducing AI.
Consider a learning sequence where students first answer a question independently, then receive teacher-approved hints, then solve a different question independently. Do not call a pilot successful merely because students enjoyed it.
Deliverables: reviewed tools and tasks, teacher practice materials, learner-facing instructions and a small set of baseline and follow-up learning checks.
Days 61–90: Evaluate and decide
Review samples of actual student explanations, error rates, equity concerns, workload and incident reports. Ask students what they can now do without assistance. Ask teachers what they had to verify. Compare findings against the original educational objective, noting sample size and limitations.
Retain uses that meet the school's requirements and show plausible benefit. Revise or discontinue others. Publish clear guidance for the next school term and schedule a further review when regulations, products or educational evidence change.
Deliverables: documented pilot decision, revised use register, family communication and the next review date.
What should school leaders measure?
A dashboard dominated by logins, prompts and platform subscriptions rewards activity rather than education. Better indicators connect safety and learning:
| Measure | What it helps answer | Important limitation |
|---|---|---|
| Independent explanations before and after a unit | Can learners demonstrate understanding without AI? | Needs comparable prompts and appropriate interpretation |
| Quality of student reasoning and source checking | Are students evaluating, rather than accepting, outputs? | Rubrics and human moderation are essential |
| Assessment authentication checks | Can teachers reasonably verify authorship and process? | No single procedure proves every case |
| Privacy or safeguarding incidents | Are risks being identified and addressed? | Low reports can also reflect under-reporting |
| Access and participation by student group | Are benefits and burdens equitably distributed? | Handle demographic data with care and permissions |
| Teacher preparation and verification time | Does a workflow reduce or redistribute workload? | Self-report is not sufficient on its own |
| Student agency and help-seeking | Can students select strategies and ask useful questions? | Surveys should be paired with observable work |
Define a baseline before rollout, and keep individual student information protected. Treat pilot results as local evidence, not proof that a particular AI system improves learning everywhere. A September 2026 University of Sydney research summary highlighted the need for stronger Australian evidence about the effects of generative AI in school settings.
From AI dependence to independent learning: the Learning Coach perspective
The central educational risk is not simply that AI can answer questions. It is that students may mistake the availability of an answer for the development of understanding. A learner can finish a task faster without becoming more capable of performing the next one.
Learning Coach advocates an approach in which adults help learners define goals, become aware of their strategies, mobilise their own resources and gradually act with greater autonomy. In an AI-enabled school, that approach means asking: “What did you think first?”, “How did you check it?”, “What would you do differently?” and, most importantly, “Can you now explain it without the tool?”
This does not mean replacing explicit instruction, curriculum expertise or specialist support. Teachers still model, teach, assess and provide feedback. A coaching stance complements those responsibilities by making student thinking visible and building independence over time.
For school leaders, the practical implication is to develop staff who can coach the learning process, rather than simply supervise tool use. Educators exploring the Learning Coach approach can learn about the training and application pathway. The page describes a pathway; the availability and scope of particular specialisations should be confirmed there before making any enrolment decision.
For further reading on learning strategies, explore the Learning Coach book catalogue and the Learning Coach education blog. These are optional brand resources, not substitutes for independent research.
Questions Australian principals are asking
Does every Australian school have to use the same AI tools?
No. Australia's national framework supplies shared principles, while approvals, procurement, privacy and implementation decisions are governed through relevant states, territories, school sectors and institutions. A tool available to one school should not be assumed approved for another.
Should schools ban student AI use until 2027?
Schools must apply any binding local restrictions and age requirements. Beyond those obligations, a blanket approach can miss the difference between acceptable support and unacceptable substitution. The stronger leadership task is to define specific permitted and prohibited uses and teach students why the distinction matters.
Can AI detectors solve academic integrity problems?
They cannot establish authorship reliably enough to replace fair assessment design and investigation. Schools should rely on their actual assessment rules, multiple sources of evidence, student explanations and proportionate procedures.
What is the most useful first investment?
Often it is staff time: a responsible governance group, assessment review, practical professional learning and careful evaluation. Procurement should follow an identified educational need and an appropriate safety review, not precede it.
How do we know whether an AI pilot worked?
Look for demonstrable progress on the learning goal, independent student performance, manageable teacher checking, equitable access and acceptable risk. Satisfaction alone is not a learning outcome.
The 2027 readiness test
An AI-ready Australian school can explain not only what technology it uses but also why it is educationally justified, what students are expected to learn, who remains accountable and how the school will recognise harm or failure. Its leaders can show an approved use-case register, teachers can explain task-level boundaries, families understand the rules, and students still have opportunities to think, struggle productively, practise and demonstrate learning independently.
That is a more enduring definition of readiness than being first to adopt the latest chatbot. The strongest schools in 2027 will not be those in which AI produces the most answers. They will be those in which students become progressively better at answering meaningful questions themselves.
Resources and references
Australian Government Department of Education — Australian Framework for Generative AI in Schools. National guidance; 2024 review endorsed June 2025.
Australian Education Research Organisation — AI in schools: considerations for school leaders and teachers. Evidence-informed considerations and limitations.
Australian Curriculum — Artificial intelligence curriculum connection. Curriculum context and AI understanding.
NSW Government — New rules for schools on AI in take-home HSC assessments. Announcement published 1 September 2026; NSW-specific changes.
NSW Department of Education — Guidelines regarding generative AI. Public guidance for NSW government schools.
Victorian Department of Education — Appropriate use of generative AI tools. Victorian policy guidance.
Safer Technologies 4 Schools — Framework v2026.1. Technology assessment framework, updated July 2026.
University of Sydney — Australia lacks evidence on GenAI in schools. Research announcement, 3 September 2026.
