Rethinking assessment for future-ready learning — with a focus on fairness, process, and reflective practice.
Why AI disrupts traditional assessment
The shift is here
As AI tools become more common, educators are asking a big question: how do we keep assessments fair and meaningful?
AI can generate essays, summaries, reports, and even creative outputs in seconds. This challenges assessments that rely heavily on traditional formats like essays — especially those that assess skills AI can now simulate, such as summarising, synthesising, or critiquing ideas.
This doesn’t mean we should panic — but it does mean we need to rethink what assessment looks like in an AI-enabled world.
Risks to integrity
- Ghostwriting via AI tools
- Over-reliance on generic outputs
- Misrepresentation of learner ability
Opportunity for innovation
Rather than banning AI, we can reimagine assessment to focus on what AI can’t easily replicate — critical thinking in context, lived experience, cultural reasoning, and human judgement. For example, we can design assessment to:
- Focus on process, not just product
- Reward reflection, originality, and context
- Integrate AI as a scaffold, not a shortcut
| Scenario: A learner uses ChatGPT to draft an initial answer. They then reflect on its limitations, rewrite the response using their own experiences or local context, and submit both versions with a short commentary on what changed and why. |
Can AI be used without undermining assessment?
AI can challenge traditional assessments — but it also opens doors for more authentic, learner-centred evaluation. When used well, AI might help scaffold learning, personalise support, or act as a ‘thinking partner’ — without replacing the learner’s voice.
Key questions to consider:
- What are you really assessing — the product, or the process?
- Can learners use AI transparently and still demonstrate original thinking?
- How might oral, project-based, or in-person assessments offer more reliable indicators of learning?
Educators across Aotearoa are already experimenting with ways to reframe assessment tasks — from reflective commentaries on AI use, to collaborative projects where process matters as much as product.
Practical steps to support integrity
You don’t need to redesign everything overnight. Start by:
- Talking openly with learners about appropriate AI use
- Including AI-related conditions in assessment briefs
- Trialling one redesigned task where AI use is expected, not hidden
These shifts help learners develop ethical judgement and digital responsibility — crucial skills for future workplaces.
Strategies for authentic assessment
Design for process
| Traditional Task | AI-Resilient Alternative |
| Write a statement | Draft a short statement, then use AI to refine it and reflect on improvements made |
| Write a report | Submit AI draft + annotated revision steps |
| Define 5 terms | Explain terms using own workplace experience |
| Answer a case study | Respond orally or via live discussion |
Want a way to map these strategies across different levels of AI use?
Leon Furze’s AI Assessment Scale offers a helpful continuum — from no AI to full AI integration. It’s already being used by tertiary institutions in Aotearoa to guide both educators and students in transparent, tiered practice.
Add reflective layers
Could learners use AI to enhance their own writing — then explain what changed, why, and how they judged its quality?
Ask students to explain:
- What did AI help with?
- What did you change, and why?
- How do you know the output was trustworthy?
Use oral or live elements
- Short verbal presentations
- 1:1 discussion of written work
- Peer review with guided AI critique prompts
These strategies can be scaled. Use 1–2 reflection prompts, quick voice notes, or peer discussion to surface thinking without adding hours of marking. Tools like professional conversations or class conversations can support this.
Upholding integrity with transparency
Talk about it early
- Clarify what kinds of AI use are allowed
- Co-create class norms or contracts
- Emphasise learning, not surveillance
Use integrity declarations
| Example: “I confirm that I used AI to generate a draft only, and have revised it significantly. I take responsibility for the final version.” |
Avoid over-reliance on detection
AI detection tools are inconsistent. They often:
- Mislabel non-native writing
- Fail to detect paraphrased content
- Penalise creativity
Instead:
Build assessments that invite authentic, situated responses — harder to fake, easier to trust.
Resources to explore and use
Templates + Tools
- University of Melbourne – Acknowledging AI Tools and Technologies
- TeacherDashboard.ai – AI Rubric Maker
Readings + Frameworks
Case examples
Want to go deeper?
- JISC – AI in Education Maturity Model (2024)
- EDUCAUSE – (Re)Designing Assessments in the Age of Generative AI
- Turnitin – AI Writing Detection Guide
This guide was part of a series, created in 2025 by Ako Aotearoa, designed to help educators explore AI in education — safely, ethically, and in ways that reflect our values in Aotearoa.
Each guide offers practical steps, real examples, and reflective tools to support your journey. Whether you’re just starting or deepening your practice, these resources are here to support your next steps — practically and purposefully.
Links were correct at the time of publish but will not be checked or updated in the future.