Refractive Design
Refractive Design is a practical approach to assignment design built around a simple principle:
Keep the decisions that matter to the learning in students’ hands.
It applies whether AI is part of the assignment or not. AI may contribute ideas, alternatives, critique, variation, feedback, or production. Students remain responsible for the decisions that matter.
What Refraction Means
Analytic refraction
An assignment can appear to be one thing: write the memo, analyze the case, solve the problem, produce the design. Refractive Design separates that assignment into the underlying design decisions that shape the learning.
Generative refraction
Once those dimensions are visible, they can be recombined in different ways to generate new assignment architectures from the same disciplinary content.
Refractive Design is both a way to examine an existing assignment and a way to generate new possibilities.
Use PRISM
PRISM provides four questions for refracting an assignment.
P — Purpose
What do you want students to learn?
What do students need to decide for themselves, and what should guide that decision?
Students might need to:
interpret or compare
evaluate
select or prioritize
diagnose
apply or establish a standard
make a tradeoff
revise
determine what is sufficient or trustworthy
The key question is not simply what students will produce, but what consequential intellectual responsibility needs to remain theirs.
R — Role
What role, if any, should AI play?
What could AI usefully do here without taking over the thinking students need to do?
AI might:
generate possibilities
critique or challenge
simulate a stakeholder
support inquiry
provide practice
assist revision
introduce variation
assist production
have no role
AI may support students’ learning, be something they are learning to use well, or both.
I — Introduce Challenge
What will require students to use, test, or extend their judgment?
What condition will prevent routine execution and make the judgment matter?
That might include:
competing alternatives
a constraint
a tradeoff
incomplete information
new evidence
a changed condition
critique or counterargument
a second case
a different stakeholder or audience
consequences of an earlier choice
The challenge can be built into the task from the beginning or introduced later.
SM — Surface Meaningful Evidence
What will let you see something meaningful about the learning?
What trace will help you interpret how students used their judgment?
Look, for example, at:
what they choose
what they reject
what changes
what remains stable
how the artifact is shaped
how a standard or threshold is applied
what they do when conditions change
the sequence of decisions
a brief decision record
a conversation or interaction
a short rationale, when one is actually needed
The goal is not to document everything students do. It is to make the moments that matter interpretable.
Start Here
Start with an assignment you already teach.
Diagnose
→ locate the consequential judgment
Refract
→ change Purpose, Role, Challenge, or Meaningful Evidence
Compare
→ generate more than one possible architecture
Redesign
→ choose the version that best supports the learning