Design Session 3 Focus
Generate possible redesign options that meet the aims of your design challenge statement, weigh the implications, and choose a direction.
Design Time
~60 minutes
Design Decisions Document
Work through the Develop section of your copy of the document to:
- Reflect on your design challenge statement you drafted previously in the Define section.
- Generate two to three redesign options.
- For inspiration on potential redesign ideas, review the Design Moves handout linked below as a starting point.
- Name the implications and trade-offs of each option. Implications are wider than cost. They include your time and workload, equity and access, privacy, new skills you would need to teach, and how learners experience the work.
- Name what you still do not know and want to test.
Design Moves Handout
This handout is a shared pool of redesign ideas, organized by the four lenses we have been using to diagnose AI vulnerability. The shared examples are by no means exhaustive or prescriptive. Instead, they are shared as starting points to spark your own ideas.
- Open the Design Moves Handout
- Select the lens in which your diagnosis suggests your learning experience is most AI-vulnerable.
- Where applicable, pull back one or two design move ideas into your own Design Decisions Document.
- The reference list at the end is an initial effort to anchor these moves in credible and relevant sources. Most sources align with multiple lenses but are listed where they first apply.
Note: The Design Moves Handout is a living document. Feel free to make a copy to use and adapt for your own needs.
Your Design Reflections
Which redesign directions are you leaning toward developing? What might be the implications or trade-offs? Jot your thoughts in your Design Decisions Document.
Jennifer’s Design Session 3: Develop
My Design Reflections:
Like everyone who worked through the live Design Sprint, I’m working through an AI-vulnerable design challenge of my own. What follows are my design session notes and reflections that align with the same design decisions you’re working through.
Direction 1: Add a synthesis matrix and light process
- Describe the direction: Keep the annotated bibliography, but require a short source-selection rationale and a synthesis matrix that maps the sources against one another and against the research question, plus one piece of reading evidence per source, such as a key passage with the student’s own commentary. This pulls from the visibility moves, adding a selection rationale and one piece of in-process evidence, with a comparison move from the tasks lens in the matrix itself.
- What does it improve? Visibility, because I see the selection reasoning, and measurement, because the matrix is tied to the student’s specific question.
- Trade-offs, and does it pass the test? More to grade. And when I picture handing this to AI, it can still draft a plausible matrix from the sources. On its own this raises the bar but does not close the gap, so it does not fully pass. It needs a personal or live anchor to become resilient.
Direction 2: Keep shorter written paper + a short 1:1
- Describe the direction: Keep a leaner written bibliography, but add a five-minute live walkthrough where the student talks me through two of their sources and why they matter to their question, and reweight so the live component carries real weight. This is the measurement move of adding a short oral walkthrough and giving it real weight, used as the attribution anchor.
- What does it improve? Visibility and measurement strongly, because attribution becomes direct, and it changes the conditions of evidence collection.
- Trade-offs, and does it pass the test? Time and scheduling, harder at scale and for asynchronous working adult students, and it will be intimidating for some. But discussing their progress, even in a brief 1:1, is a great formative assessment to also help answer their questions. Double-check it’s ok with the department. Assuming so, this one passes.
Direction 3: Reframe the task toward judgment
- Describe the direction: Ask students to annotate sources in explicit dialogue with their own question and with each other, for example “which two of these would you build on and why, and which would you set aside,” plus a short reflection on where they used AI and where it was wrong or insufficient. This combines the task of anchoring the work to the learner’s own question with the structured reflection on AI use, which folds AI literacy directly into the task.
- What does it improve? Tasks, because it requires judgment tied to their specific question, and measurement, and it folds AI literacy in directly.
- Trade-offs, and does it pass the test? A larger redesign and rubric rework, and students will need scaffolding to do real synthesis. Picturing the AI attempt, AI could produce a generic “which to build on,” but not one grounded in the student’s actual evolving question and reading, and the reflection on their own AI use is inherently personal. It passes more than Direction 1, though less directly than the live component in Direction 2.
Note regarding AI-Vulnerability Test of these options: For each direction, I am not just asserting it would 100% pass the AI-vulnerability test. I think it will help, but I’ll hand the redesigned task to Claude as I develop and test the redesign to see what it could still do.
My Preferred Decision
- Which to pursue first, or what to keep alive? I want to pursue a combination first: Direction 1 as the written core, with Direction 2’s short oral walkthrough as the attribution anchor. I am keeping Direction 3 alive as a possible next iteration, and what a test would need to show before I adopt it is whether students can do real source-to-source synthesis in writing without the live prompt.
- What feels strongest? Picturing the AI attempt on each direction made the decision for me. The written layer alone is still only partially completeable by AI, and the oral walkthrough is the piece that makes authorship observable. The live component is load-bearing, which I would not have seen if I had only asserted that the matrix “would pass.”
- Uncertainties into Deliver: The logistics of the oral component at scale, and how to handle asynchronous students fairly.
Next: Design Session 4 | Plan to Deliver
Previous: Design Session 2 | Back to Design Sprint On-Demand