Subject playbook

AI for HSC / Stage 6 Teachers

The senior-years playbook: use AI for exam prep, marking to the standards and feedback at pace — while keeping authorship, accuracy and HSC integrity beyond doubt.

Modules

Each module: clear learning outcomes → short, accurate senior-secondary input → a hands-on activity using the Lessio generator → interactive knowledge checks. Built for the HSC year, with marking-to-the-standards and integrity as the spine. Mapped to the Australian Professional Standards for Teachers.

Click a module to read it.

1. 1

   Where AI helps in the HSC year — and where it must not

   The senior-years map: where AI genuinely buys you time under HSC pressure (exam prep, scaffolding, feedback at pace) and the high-stakes limits where it must stay out.~45 min

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

   • Identify the senior-secondary tasks where AI genuinely saves time, and the high-stakes tasks where it must not lead.
   • Apply teacher-in-the-loop at HSC stakes — where a verification slip costs a student marks or an award.
   • Recognise the three HSC-specific failure modes (fabricated content, band-descriptor misalignment, integrity) before they reach a student.

   Standards2.6 Information and Communication Technology (ICT)6.2 Engage in professional learning7.1 Meet professional ethics and responsibilities

   This playbook assumes the flagship — it doesn't repeat it

   You've done (or will do) Teaching with AI: Ethical & Effective Practice — the general ethics, the three-layer policy stack (national Framework / NSW DoE / NESA), the hard line on student data, and the RICE prompt structure. This playbook does not re-teach those. It takes them as given and goes straight to the senior-secondary edge, where the stakes — and the temptation to cut corners with AI — are at their highest.

   The principle is unchanged but the consequences are sharper: in the HSC year, you remain the professional and the author, and a verification slip you'd shrug off in Year 8 can cost a senior student marks, or an award.

   Why senior teaching is the hard case

   The HSC year is a pressure cooker: heavy school-based assessment, trials, marking turnarounds, and an external exam whose results — reported in performance bands against band descriptors — feed an ATAR. Under that pressure, the shortcuts AI offers look most tempting precisely where they're most dangerous: marking a class set overnight, handing out a generated "Band 6 response", letting a tool do the judging. This module draws the map before you reach for the tool.

   Where AI genuinely helps a senior teacher

   Used as a drafting assistant you verify, AI earns its place in the senior year:

   • Exam-style practice at volume — generating and varying exam-style questions for a topic so students get fresh retrieval and exam-condition practice (Module 2).
   • Marking guidelines aligned to the standards — drafting a marking guideline structured around the NESA performance-band descriptors, which you then align and verify (Module 2).
   • Exemplar analysis — articulating what separates a Band 5 from a Band 6 response, as a teaching tool (Module 3).
   • Feedback frames at pace — structuring strength / priority / next-step feedback on extended responses for a whole class, fast, against your criteria (Module 3).
   • Revision and retrieval resources — summaries, glossaries, spaced-retrieval sets, study scaffolds, differentiated for your cohort.

   Where it must not lead — the high-stakes limits

   Task	Why AI must not lead
   Awarding the actual marks / the final judgement	Marks against the NESA standards are your professional judgement; they feed an ATAR.
   Handing students a "model answer" unverified	AI confidently invents wrong content — an unverified "Band 6" answer mis-teaches a whole class.
   Doing the marking to the AI's own standard	The AI doesn't know the band descriptors; its "marks" are not NESA-aligned.
   Anything presented as a student's own work	AI-assisted work passed off as the student's is malpractice (Module 4).

   The three HSC-specific failure modes — name them now

   This playbook returns to these throughout. Hold them from the start:

   1. Fabricated or distorted subject content. AI invents plausible, wrong "model answers" and subject detail with total confidence. Never hand students an unverified AI exemplar.
   2. Band-descriptor misalignment. AI "marking" reflects generic notions of quality, not the NESA performance-band descriptors for your course. Its judgement ≠ the standard.
   3. Integrity at the highest stakes. Submitted assessment tasks and major works carry the student's "all my own work" declaration. AI-assisted work presented as a student's own is malpractice that can jeopardise their HSC.

   And a quieter fourth: over-reliance in a high-pressure year is its own risk — to your students' independence and to your own expertise. The time AI gives back should go into teaching and feedback, not into shipping more unchecked material.

   Activity — map your HSC year (10 min)

   List the recurring tasks of your Stage 6 year — writing practice questions, marking, feedback, exemplars, revision resources, the final grade. Mark each AI can draft (I verify), AI assists, I finish, or stays fully mine. Then open the Lessio generator and produce a short set of exam-style questions for one senior topic you teach — note how fast the starting point arrived, and write down the one thing you'd verify before a student saw it. That verification is your job, and at HSC stakes it isn't optional.

   Knowledge check

   1Name the three HSC-specific failure modes this playbook keeps returning to.

   Show answer

   2A colleague wants to mark a full class set of trial responses overnight by pasting them into a chatbot and using its marks. What's the senior-years problem?

   Show answer

   3Why is 'teacher-in-the-loop' even more demanding in the HSC year than in junior years?

   Show answer

   Mark this module complete
2. 2

   Exam-style questions, past papers & marking to the standards

   Generating and curating exam-style questions, then drafting marking guidelines aligned to the NESA performance-band descriptors — and the 'mark like a marker' discipline that keeps the judgement NESA's, not the AI's.~55 min

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

   • Generate and curate exam-style questions and variations that match the demand of your course, then verify them.
   • Draft a marking guideline aligned to the NESA performance-band descriptors and check it against the actual standard.
   • Apply the 'mark like a marker' discipline so judgements stay consistent, comparable and NESA-aligned.

   Standards2.3 Curriculum, assessment and reporting5.3 Make consistent and comparable judgements2.1 Content and teaching strategies of the teaching area

   The senior teacher's highest-value use — done safely

   Two of the most time-consuming jobs of the HSC year are writing exam-style questions and building marking guidelines. AI can draft both fast. Whether that's a help or a hazard comes down to one thing: do you treat the draft as a starting point you verify against the syllabus and the band descriptors, or as a finished product? This module makes it the former.

   Generating & curating exam-style questions

   A good prompt (RICE, from the flagship — role, intent, constraints, examples) gets you a usable bank:

   • Match the demand, not just the topic. Senior questions live at specific levels of demand and use the syllabus's directive verbs (analyse, evaluate, assess, justify, explain). Ask for a spread across command terms and mark values, not ten recall questions.
   • Generate variations. One stem, several versions — different data, contexts or numbers — so a class can't simply share answers and students get fresh retrieval.
   • Curate, don't just accept. AI invents plausible-but-wrong subject content. Every generated question and answer is verified for accuracy and syllabus fit before it goes near a student — this is failure mode #1 from Module 1.
   • Past papers stay the gold standard. AI complements official NESA past papers and your faculty bank; it doesn't replace them. Use it to multiply practice around the real thing.

   A generated question that's subtly off-syllabus, or whose "answer" is wrong, teaches the wrong thing to a whole class. Curation is the work.

   Marking guidelines aligned to the band descriptors

   This is where senior marking earns its rigour. NESA reports results in performance bands — Bands 1–6 for most courses, Bands E1–E4 for Extension courses — described by band descriptors that set out what performance at each band looks like. A marking guideline must point at those standards.

   AI can draft a marking guideline structured around the descriptors — but it does not know them. Your job:

   1. Anchor to the real descriptors. Give the AI your course's band descriptors / marking-guideline structure as the reference (de-identified — no student work needed to build a guide).
   2. Align, then verify. Check that the draft's criteria and mark allocations actually reflect the NESA standard for the band, not a generic rubric the model invented. This is failure mode #2 — band-descriptor misalignment.
   3. Keep the holistic judgement. Senior marking is often holistic against a band, not a mechanical tick-list; the guideline supports your judgement, it doesn't replace it.

   The "mark like a marker" discipline

   Marking for the HSC is about consistent and comparable judgement (APST 5.3) — the same response earns the same mark whoever marks it, and the mark means what NESA says it means. So:

   • Mark to the descriptor, never to the AI. The AI's sense of "good" is not the standard; the band descriptor is.
   • Use AI to support consistency, not to outsource it. It can help you articulate why a response sits in one band rather than the next, draft annotations, or sanity-check your reasoning against the criteria — with you making every call.
   • Cross-check and moderate as a faculty. AI-assisted annotations are an input to faculty moderation, not a substitute for it. Comparable judgement is built between colleagues, not by a chatbot.
   • De-identify and protect data. As ever (flagship, Module 2): don't paste identifiable student work into a general tool; use a de-identified response or your school's approved, secured environment.

   Where Lessio fits

   Lessio generates NSW-grounded assessments with marking guidelines and graduated questions. For a senior teacher that means: draft an exam-style task and a marking guideline for your subject in minutes, then do the professional work — verify the subject accuracy and align the guideline to your course's band descriptors before it's used. Lessio gives you the structured starting point; the alignment to the standard is yours.

   Activity — generate a question + guideline, then align it (15 min)

   In Lessio, generate an exam-style question with a marking guideline for a senior topic you teach. Then do two things: (1) verify the question and any model answer for subject accuracy against the syllabus, and (2) align the marking guideline to your course's performance-band descriptors — adjust at least one criterion so it reflects the NESA standard rather than a generic one. The adjustments you make are the visible proof of marking to the standard.

   Knowledge check

   1An AI-drafted marking guideline reads well but its criteria don't match your course's band descriptors. What's the issue and what do you do?

   Show answer

   2How are senior results reported, and why does that matter for how you build a marking guideline?

   Show answer

   3Does AI replace official NESA past papers for exam-style practice?

   Show answer

   Mark this module complete
3. 3

   Scaffolding Band 5/6 responses & giving targeted feedback at pace

   Using AI to analyse exemplars (what separates a Band 5 from a Band 6), build feedback frames for extended responses at volume, and scaffold senior writing — without writing it for the student.~50 min

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

   • Use AI to analyse exemplars and articulate what distinguishes a Band 5 from a Band 6 response.
   • Build targeted feedback frames for extended responses at volume — against your criteria, without writing the response for the student.
   • Scaffold senior writing while keeping the thinking, and the words, the student's own.

   Standards5.1 Assess student learning5.2 Provide feedback to students on their learning5.4 Interpret student data

   The pace problem in senior feedback

   Extended responses are the heart of senior assessment — and the slowest thing to give feedback on. A class set of essays, source analyses or extended-response questions can swallow a weekend. AI can help you give consistent, targeted feedback faster. The discipline that keeps it honest: AI helps you frame and target feedback; it never writes the response for the student.

   Exemplar analysis — teaching what Band 5 vs Band 6 looks like

   One of the most powerful senior teaching moves is showing students concretely what separates a strong response from a top one. AI is genuinely useful here — as an analysis tool for you:

   • Take a de-identified, verified exemplar (or two at different bands) and ask AI to articulate what the higher-band response does that the other doesn't — sophistication of argument, integration of evidence, control of form, conceptual understanding.
   • Map that articulation back to your course's band descriptors so the distinction is the NESA one, not a generic one.
   • Turn it into student-facing language: "to move from Band 5 to Band 6 here, you need to…".

   The exemplar you analyse must be real and verified. Never generate a fake "Band 6 response" and hand it to students as a model — failure mode #1. A confident, fluent, wrong exemplar mis-teaches the whole class.

   Feedback frames at volume — targeted, not generic

   For a class set, AI can draft a feedback frame against your success criteria so every student gets specific, structured feedback fast:

   • Feed it the criteria / marking guideline and a de-identified response; ask for one genuine strength, one priority for improvement tied to a criterion, and one specific, actionable next step.
   • Keep it specific to the response — instruct it not to invent quotes or content that isn't there.
   • You personalise and verify every frame. The frame is a scaffold for your judgement, not a verdict; you know the student and the trajectory, the AI doesn't (this is interpreting your data, APST 5.4 — patterns across the set tell you what to reteach).

   This is augmentation: AI structures the feedback at pace; you make it true, kind and right for the actual student.

   Scaffolding Band 5/6 writing — without doing the writing

   The line that protects both learning and integrity: scaffold the thinking, never supply the response.

   • ✅ Generate planning scaffolds, paragraph structures, question-deconstruction frames, sentence-level "how to lift this" guidance, and worked analysis of technique — tools that build the student's own capability.
   • ✅ Build revision and retrieval resources that strengthen the knowledge a Band 5/6 response draws on.
   • ❌ Don't generate a finished extended response for a student to lean on, copy or submit. A student who hands in AI-written work — even "just to improve it" — is heading into the integrity problem of Module 4, and isn't learning to write at band.

   Where Lessio fits

   Use Lessio's graduated questions and marking guidelines to build the scaffolds and retrieval practice that move students toward Band 5/6 — verified against your syllabus — and use its marking-guideline structure as the criteria backbone for your feedback frames. The frames and the exemplar analysis are yours to verify and personalise.

   Activity — a feedback frame, then an exemplar move (15 min)

   Take a de-identified senior extended response and your success criteria. In your tool, generate a feedback frame (strength / priority / next step). Personalise and correct it for the (de-identified) student. Then take a verified exemplar and draft one student-facing "to move from Band 5 to Band 6 here, you need to…" statement, mapped to your band descriptors. Note where your professional judgement overrode the AI's framing.

   Knowledge check

   1Why must you never generate a fake 'Band 6 response' and hand it to students as a model?

   Show answer

   2What's the line that keeps AI scaffolding on the right side of learning and integrity?

   Show answer

   3How does AI help give targeted feedback at pace without taking over the judgement?

   Show answer

   Mark this module complete
4. 4

   HSC integrity & authorship — the highest-stakes version

   NESA's HSC academic-integrity rules and malpractice, the student's 'all my own work' obligation, major works and projects, and designing AI-resilient senior assessment by design — not by 'AI detectors'.~50 min

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

   • State NESA's HSC integrity expectations and what counts as malpractice, including the 'all my own work' obligation.
   • Apply integrity rigour to major works and projects, where the authorship stakes are highest.
   • Design AI-resilient senior assessment by making process and authorship visible — not by relying on 'AI detectors'.

   Standards7.1 Meet professional ethics and responsibilities7.2 Comply with legislative, administrative and organisational requirements2.3 Curriculum, assessment and reporting

   Why integrity is the spine of this module

   Everything in the senior year resolves to one declaration: the student's work is their own. The general principle was in the flagship; here it is at full HSC stakes, because a result determines an ATAR and because senior assessment carries formal integrity obligations that earlier stages don't.

   The "all my own work" obligation

   HSC students formally declare that submitted assessment work is their own. NESA's academic-integrity expectations make the categories of malpractice clear:

   • Plagiarism — presenting someone else's (or something else's) words or ideas as your own.
   • Collusion — improperly working with others, or submitting work that isn't genuinely the student's.
   • Misrepresentation — claiming work as your own that isn't.

   AI-assisted work presented as a student's own is malpractice. A student using AI to draft, write or substantially produce work they then submit as theirs breaches the "all my own work" obligation — and at HSC stakes that can jeopardise their result or award. This is failure mode #3, and it's the one with the heaviest consequences.

   Schools decide, task by task, whether and how AI may be used (NESA leaves this to schools). The teacher's job is to make that boundary explicit for every senior task and to design the task so the boundary holds.

   Major works and projects — the highest authorship stakes

   Many HSC courses include major works, projects or performances developed over months. Their integrity is assured the way NESA has long assured project authorship: monitoring authorship over time, not policing the final artefact.

   • Build in checkpoints: process diaries, logbooks, drafts, supervisor meetings, in-progress sightings.
   • Talk to the student about their decisions and thinking — authentic authorship shows up in conversation.
   • Make clear where AI may and may not feature, and document the student's own contribution.

   A major work is where "all my own work" matters most and where authorship-by-process protects the honest student.

   Design AI-resilient assessment — by design, not by detection

   NESA's guidance is to assure authorship by design, not by "AI detectors" — which are unreliable and can wrongly accuse. AI itself can help you redesign senior tasks to be resilient:

   Less resilient	More resilient (authorship visible)
   Take-home essay, single submission	Drafting at checkpoints + an in-class final or viva on the same task
   Generic prompt answerable from anywhere	Prompt tied to this class's case study, fieldwork, text or data
   Product only assessed	Process assessed — plan, drafts, reflection, logbook
   One-shot extended response	Staged: planning sighted, writing under supervision, oral defence of choices

   Use AI to generate these resilient variants and the in-class/viva components — then verify them. You're using the tool to protect integrity, not to chase breaches after the fact.

   Modelling integrity to senior students

   Senior students will use AI. Teach them where the line is for each task, why the "all my own work" declaration matters, and how to use AI with integrity — as a tutor or revision aid they critique and disclose, not a ghost-writer. Your own honest, disclosed practice (flagship, Module 5) sets the standard you're asking them to meet.

   Where Lessio fits

   Use Lessio to draft AI-resilient senior tasks — assessments tied to your class's context, with marking guidelines and the in-class or staged components that make authorship visible — then verify and align them. Designing for integrity is faster when the task generator works with you.

   Activity — make one senior task AI-resilient (15 min)

   Take a real senior assessment task. Use Lessio (or your tool) to redesign it for visible authorship: add a process or checkpoint component, tie the prompt to your class's context, and add an in-class or viva element — then verify the redesign against the syllabus and your faculty's AI position. Write the one sentence you'd tell students about where AI may and may not be used on this task.

   Knowledge check

   1Is a student using AI to write an essay they submit as their own a problem under NESA's rules? Why?

   Show answer

   2How does NESA expect you to assure authorship — and how should you NOT?

   Show answer

   3Why do major works and projects carry the highest authorship stakes, and what protects the honest student?

   Show answer

   Mark this module complete
5. 5

   Capstone — build an HSC-aligned task or practice exam, then critique it

   Build a real, syllabus-aligned senior task or practice exam with marking guidelines, critique it against the band descriptors and the integrity rules, self-assess against the HSC Ethical-Use Checklist, and log it as PD.~50 min

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

   • Build a syllabus-aligned senior assessment task or practice exam with marking guidelines using AI, end to end.
   • Critique and improve it against the syllabus, the NESA band descriptors and the HSC integrity rules.
   • Self-assess against the HSC Ethical-Use Checklist, reflect, and record the hours as PD.

   Standards2.3 Curriculum, assessment and reporting5.3 Make consistent and comparable judgements6.2 Engage in professional learning

   The task — a real, defensible senior assessment

   Choose a senior topic you'll teach or assess this term. Using Lessio, build and then critique:

   1. An exam-style task or short practice exam for the topic — questions that match the demand and command terms of your course.
   2. A marking guideline for it, structured to align with your course's performance-band descriptors.
   3. The integrity design — how authorship is made visible (a checkpoint, a class-specific element, an in-class or viva component).

   Then bring your professional judgement to bear: verify the subject accuracy of every question and model answer against the syllabus; align the marking guideline to the band descriptors (failure mode #2); make the task AI-resilient so it upholds the "all my own work" obligation (failure mode #3); and check it's valid (it measures the outcome) and fair (accessible, with reasonable adjustments under the Disability Standards for Education 2005).

   What good looks like

   A senior task and marking guideline you'd actually set and defend at a faculty meeting — drafted by AI, unmistakably verified, aligned and owned by you. Your edits — the corrected content, the descriptor-aligned criteria, the authorship safeguards — are the evidence of marking-to-the-standard and integrity-by-design in practice.

   Self-assessment — the HSC Ethical-Use Checklist

   Run your capstone against all five checklist items on this page. Every box should be honestly tickable — especially no unverified AI 'model answer' given to students, marking aligned to the NESA band descriptors not the AI's, and HSC integrity rules upheld. If one isn't, fix the artefact — that is the learning.

   Reflection — write a short response

   • What did AI genuinely save you time on in this senior task, and what did you have to fix or verify?
   • Where did aligning to the band descriptors change the marking guideline?
   • What did you change to make the task AI-resilient and protect the "all my own work" obligation?
   • One rule you'll keep for using AI in your HSC year from now on.

   Log it as professional learning

   This module is your assessment: a complete, critiqued senior artefact plus your ethical-use reflection — keep it as evidence of practice. Since NESA removed the Accredited/Elective PD distinction in 2024, Standards-relevant learning like this counts toward your 100 maintenance hours — record it in your eTAMS PD log against the Standards it addresses (especially Standard 5 — assess, feedback, consistent judgement — plus 2.3, 4.5 and 7). Your faculty can also run this playbook as part of its professional-learning plan or a senior-assessment twilight.

   Knowledge check

   1What turns an AI-generated senior task into defensible professional work?

   Show answer

   2Which Ethical-Use Checklist items most define a senior capstone, and why?

   Show answer

   3How can this playbook count toward your NESA professional-development hours?

   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.

You are an experienced NSW Stage 6 [subject] teacher. Write six exam-style questions on [topic] for [de-identified course/year], spread across the relevant command terms (e.g. explain, analyse, evaluate) and mark values, matching the demand of the course. Include an answer guide for each. Use Australian context and conventions. At the end, flag anything you are not certain is subject-accurate or syllabus-aligned so I can verify it.

Then I will check every question and answer against the syllabus and our past papers before any student sees it.

You are a NSW Stage 6 [subject] teacher. Using the performance-band descriptors I paste below as the reference, draft a marking guideline for this extended-response question [paste], with criteria and mark allocations that reflect what each band requires. Keep it faithful to the descriptors, not a generic rubric.

[paste band descriptors / course marking-guideline structure]

Then I will align and verify it against the actual NESA band descriptors for the course, and mark to the descriptor — not to your output.

You are a NSW Stage 6 [subject] teacher. Here are two de-identified, verified responses to the same question at different bands [paste]. Using our course's band descriptors as the reference, articulate precisely what the higher-band response does that the other does not (argument, evidence, control of form, conceptual understanding), then phrase it as a student-facing 'to move from Band 5 to Band 6 here, you need to…' statement.

I confirm both exemplars are real and verified — do not invent or 'improve' them. I will check the analysis against the descriptors before sharing it.

You are a NSW Stage 6 [subject] teacher. Using these success criteria [paste], write a feedback frame for this de-identified extended response [paste]: one genuine strength, one priority for improvement tied to a specific criterion, and one specific, actionable next step. Warm, specific, jargon-free, addressed to the student. Do not invent quotes or content that isn't in the response.

I will personalise and verify every frame for the actual student before it goes back.

You are a NSW Stage 6 [subject] teacher. Write a retrieval-practice set of ten questions on [topic] for [de-identified course/year], graduated from recall to exam-style application, with an answer key. Plain English, Australian context. Flag any question where the answer could be ambiguous or where you're unsure of subject accuracy.

I will verify the questions and answers against the syllabus before use.

You are a NSW Stage 6 [subject] teacher. Create a one-page revision resource for [topic] for [de-identified course/year] that scaffolds how to plan and structure a Band 5/6 extended response: a question-deconstruction frame, the key knowledge to draw on, and a paragraph/structure scaffold. Do NOT write a full model response — build the student's own capability. Australian context, plain English.

I will verify it against the syllabus and confirm it scaffolds the thinking rather than supplying the answer.

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