Designing Inclusive Assessments in the Age of AI

AI doesn’t just change how students complete assessments — it changes who is included, who is disadvantaged, and whose voice becomes invisible if we’re not intentional.

This section offers inclusive, AI-aware design moves that support diverse learners without diluting academic challenge.

1. Assess the Human Layer, Not the AI Layer

Shift marks toward:

• decision-making

• critique

• explanation

• context

• reflection

• localisation

These are the parts that AI cannot authentically produce.

Try:

Instead of “Write 800 words on X,” shift to:

“Use any tools you like to explore X. Then explain:

• what you agreed with,

• what you disagreed with, and

• what you added from your own context.”

Outcome:

Students with AI access don’t gain an unfair advantage; students without it aren’t disadvantaged.

2. Use Multi-Modal Expression as a Default Option

Let learners demonstrate understanding through:

• audio explanation

• short video

• annotated screenshots

• mindmaps

• paragraph + voice note combos

Why it matters:

Reduces barriers for ESOL, neurodivergent learners, and anxious writers.

3. Require Evidence of Process — Not Just Product

Ask for:

• prompt attempts

• drafts

• AI comparisons

• revision notes

• “what I changed and why” blurbs

• mistake analysis

This protects integrity and reveals thinking.

4. Ground Tasks in Aotearoa Contexts

AI is weakest when tasks are:

• local

• specific

• cultural

• personal

• contextual

• relational

Examples:

• “Apply this idea to your whānau, workplace, or community.”

• “Explain using an example from Aotearoa.”

• “How would this concept work in your rohe (region)?”

Outcome:

Learners must bring themselves into the task.

5. Build Choice Into Assessment Pathways

Choice increases engagement and reduces inequity.

Offer:

• AI-supported pathway (critique + curate + reflect)

• AI-free pathway (build + revise + reflect)

Both aligned to the same learning outcomes.

6. Scaffold AI Literacy Directly Into the Assessment

Examples:

• “Highlight one moment AI was wrong — and correct it.”

• “Describe how AI shaped or challenged your thinking.”

• “What cultural perspective did AI miss?”

• “What did you remove because it didn’t fit your community?”

Outcome:

Critical AI literacy becomes part of the assessment, not a bolt-on.

7. Design for Transparency, Not Policing

Instead of detection software, use:

• reflective disclosures

• voice explanations

• annotated drafts

• face-to-face kōrero

What you’re assessing:

authenticity + understanding + reasoning.

Not surveillance.

Kaupapa Māori Lens — Aromatawai that Uplifts Mana

A culturally anchored layer that enhances LP3’s “Mana Deep Dive” but avoids repetition.

🪶 1. Mana Motuhake as the Assessment Anchor

Design tasks that preserve the learner’s voice, not overwrite it.

Prompt:

“What part of this mahi could only have come from you or your whānau?”

This keeps identity central and prevents AI homogenisation.

🪶 2. Whakapapa as Evidence of Learning

Instead of asking, “Did you use AI?”

Ask, “What is the whakapapa of your ideas?”

Encourage mapping:

• people

• places

• experiences

• texts

• conversations

• AI outputs

Whakapapa restores relational depth.

🪶 3. Whanaungatanga in Assessment Conversations

Use collaborative interpretation:

“What did our class agree the AI missed?”

“What did your tīpuna teach you that AI wouldn’t know?”

This centres relational learning over tool dependence.

🪶 4. Tapu + Taonga

Some assessments are human-only:

• whakapapa

• lived experience

• spiritual reflections

• cultural practices

Encourage learners to decide:

“Is this something I should give to a machine?”

This builds ethical discernment.