
Rebuild your grading criteria to reward reasoning over correctness
You'll have updated rubrics or marking guides that prioritize how students explain their thinking—work that makes AI-generated answers obviously inadequate and human thinking obviously essential.
What you'll make
Reasoning-First Assessment Rubric
- Cognitive Checkpoints: the thinking moves you're actually grading
- Evidence Markers: what authentic reasoning looks like in student work
- AI Failure Points: where shortcuts become obvious
- Implementation Notes: how to use this rubric consistently
Write your rubric as a working document with specific criteria and scoring guides you'd actually hand to students
The 4 parts
Each part ends with a check you can see.
Audit Your Current Grading
Identify where your existing rubrics reward outputs that AI can easily produce.
Done when
A ranked list of your current criteria from most to least AI-vulnerable, with specific examples of what AI produces for each.
Design Cognitive Checkpoints
Create assessment criteria that target the thinking process itself, not just its outputs.
Done when
5-7 new rubric criteria that explicitly require students to expose their reasoning process, with examples of what high/low performance looks like for each.
Test Against Student Work
Apply your new criteria to actual student responses and refine based on what you discover.
Done when
Evidence that your new criteria successfully distinguish between authentic reasoning and AI-assisted shortcuts across at least 6 student samples.
Finalize Your Reasoning Rubric
Complete your tested rubric with implementation guidance for consistent use.
Done when
A complete rubric document ready for immediate classroom use, with scoring guides that prioritize reasoning over correctness.
Capstone: you put together your Reasoning-First Assessment Rubric.
Where this helps
An elective: a single lesson that sits outside rtwo's job certificates and doesn't count toward one.
About this lesson
What you'll be able to do
- Design rubrics that make human reasoning essential and AI shortcuts insufficient
- Test your new criteria on actual student work to identify where AI fails to demonstrate understanding
- Have a systematic way to distinguish authentic thinking from AI-assisted surface completion
Who it's for
- K-12 teachers who've noticed AI making their current assessments meaningless
- Educators redesigning assignments to stay ahead of AI capabilities
- Department heads rebuilding assessment standards that actually measure learning
What it won't cover
- Detecting AI-generated work after the fact
- Teaching students about AI bias or responsible use
- Completely AI-proof assignments — the goal is AI-inadequate criteria, not AI-impossible tasks
Skills you'll practise
- Cognitive task analysis for assessment design
- Evidence gap identification in student reasoning
- Iterative rubric testing with real samples
- Motivation-first assignment sequencing