Context before output
AI responses are grounded in the active design step and the cumulative design context.
Research model
Study the reasoning,
not just the output.
InSTEMAI is designed to make learning-design decisions, AI use, evidence, and revision visible enough to examine — without turning the participant experience into a research dashboard.
Design
reasoning
reasoning
Explore the model
Design reasoning
What changes in the decisions?
The system keeps design choices connected across the process so researchers can examine how ideas become more specific, justified, and coherent over time.
Visible unitDesign decisions
Across timeEarlier → revised thinking
Key questionWhat changed, and why?
Design principles
Research-ready underneath.
Participant-first on the surface.
The research architecture is intentional, but it should never make the learning experience feel like instrumentation.
Evidence before claims
The system distinguishes known information, assumptions, and points that still require verification.
Revision stays visible
Snapshots, feedback, and revisions preserve the development of the artifact instead of hiding it.
Human judgment stays central
AI can question, compare, critique, and suggest. Consequential design choices remain the designer’s.
Transparency
What the product makes visible — and what it keeps out of the way.
Participant-facing simplicity is a design requirement, not an absence of rigor.
Visible to the participant
- The next meaningful design decision
- Why that decision matters
- AI questions, examples, critique, and STEM-specific guidance
- The evolving learning-design artifact
- Clear progress through the design process
Kept out of the participant workflow
- Internal model-comparison labels
- Research telemetry and implementation details
- Developer-facing configuration state
- Artificial scores presented as automated grades
- System architecture that does not help the learner act
Research-informed. Still designed to be used.
Open the Studio