NTLSN · Crash Course · Integrity

Academic integrity in the age of AI — a crash course

Generative AI hasn’t broken integrity — it has changed where the work of upholding it sits. Four short lessons on designing for integrity, then a self-check.

The one thing to remember: integrity is now about designing for it, not just policing it. Be clear with students about acceptable AI use, and build assessment that values the process, not only the product.

4 lessons~10 min read1 self-checkGrounded in educative-integrity good practice (TEQSA & the sector)

The lessons

Be educative, not just punitiveIntegrity as something students learn

›

Detection-and-punishment alone doesn’t teach students what integrity is or why it matters. An educative approach treats integrity as a capability students develop — and one we share responsibility for.

Teach what counts as acceptable practice, including when and how AI can be used.
Respond to lapses proportionately, with a developmental as well as a disciplinary lens.
Frame integrity around learning and professional standards, not just rule-following.

Grounded in

Educative approaches to academic integrity (TEQSA & the sector)
Whole-of-institution integrity good practice

Design AI-resilient assessmentAuthentic, programmatic, hard to outsource

›

Where the tool can produce the artefact in seconds, the artefact alone is a weak signal of learning. Assessment reform — not surveillance — is the durable response.

Assess authentic, contextualised tasks tied to your students, course and discipline.
Value the process: drafts, reflections, vivas, in-class or oral checkpoints.
Think programmatically — secured points of assurance across a program, not every task.

Be transparent about AI useClear expectations and declarations

›

Most confusion isn’t defiance — it’s ambiguity. Students need to know, per task, what is and isn’t allowed, and how to disclose what they did.

State course- and task-level AI expectations explicitly, not just in policy.
Ask students to declare how they used AI, and acknowledge it as they would any source.
Keep expectations consistent within a course so the rules aren’t a guessing game.

Support students so they don’t need to cheatWorkload, belonging, help-seeking

›

Misconduct rises where students feel overloaded, isolated, or unable to ask for help. Conditions matter as much as rules.

Smooth workload and deadline bunching where you can — pressure drives shortcuts.
Build belonging and make help-seeking normal, early and low-stakes.
Signpost academic skills, language and wellbeing support before the crunch, not after.

◇ Bring it together — from the NTLSN commons

Browse “Academic integrity”Assessment resources Teaching with AI Frameworks & policy

Before your next assessment design — a quick self-check

✓

I treat integrity as something students learn, not just something I police.

✓

My responses to lapses are proportionate and developmental, not only punitive.

✓

My assessment values the process — drafts, reflection, oral or in-class checkpoints.

✓

I assure learning programmatically, not by trying to lock down every task.

✓

I state acceptable AI use per task and ask students to declare what they did.

✓

I reduce pressure and make help-seeking normal so students don’t feel pushed to cheat.

Source & attribution. Curated from academic-integrity good practice — including widely-used sector work on educative approaches and assessment reform (in Australia, the regulator TEQSA emphasises both) and the broader contract-cheating research base — indexed by the NTLSN commons. Practitioner synthesis, not original research; always follow your own institution’s academic-integrity policy.

A free NTLSN crash course · see them all at /crash-courses.html. Runs in your browser; nothing is stored or sent. · ← Back to NTLSN