Tomorrow Ready Resources → Evaluating AI Output → NCEA Integrity by Design
[S] Secondary Evaluating AI Output

NCEA Integrity by Design

Authenticity rests on reasoning you can see and hear, not on tool policing. The most workable NCEA path is to design tasks where genuine thinking is hard to fake because it is visible and traceable.
Classroom Strategy

Add a matchup table requirement to one internal practice task. Mark the table and the explanation, not just the final paragraph.

  • Set a debatable statement connected to the standard
  • Students generate two candidate claims
  • Students find evidence within class-approved materials for each claim
  • Students complete a matchup table showing how the same evidence supports each claim differently
  • Students write one paragraph using the chosen claim with a rejection note: "I rejected Claim B because..."
The In-Class Pivot

The matchup table becomes more powerful when paired with a live variable change in class. After students complete the table, change one element: a new quote, a different data point, or a shifted scenario. Students write a short paragraph: what changes now, and why?

This seals the integrity move. Copy-paste finished work cannot adapt to the new variable. Genuine understanding can. The pivot is the design that makes the difference visible without any detection tool required.

Curated Source Condition

Build the task so students must compare within a shared evidence set: class texts, notes, and approved articles. Improves equity, integrity, and manageability simultaneously. Reduces the risk of students using unknown sources.

Optional Mini-Conference

After the matchup table, ask the student to explain their choice using the table, not the paragraph. Two minutes. One question: why did you choose this claim over the other? The answer reveals ownership of the reasoning.

Decision Checkpoint

Does the comparison artefact actually reflect the learning intention: interpretation, evidence use, and reasoning?

Governance Reminder

Are prompts and texts locally grounded and respectful, with space for multiple perspectives? Students use only class materials, not personal data or identifying contexts. Cultural safety applies to the evidence set as much as to the task design.

Print or save this resource as a PDF using your browser.

More resources in this theme

Browse all Evaluating AI Output resources →
← Previous: The Oversimplification Audit ← Back to Evaluating AI Output