Make the distribution of epistemic responsibility an explicit object of instruction, including think-first routines before consulting the AI¶

Strategy · All strategies
Evidence · no claims cited

Description¶

The authors suggest teachers make explicit which parts of a task students want AI to support, which processes they are expected to practice themselves, and how to recognize when assistance has begun to substitute for those processes. One possibility is to "introduce think-first routines in which students articulate an initial interpretation, strategy, or uncertainty before consulting the AI", then use the model to extend, challenge, or critique their reasoning. Students could also compare different ways of sequencing AI assistance and reflect on what they understood and could subsequently do independently.

Design Implications¶

Context¶

Requirements¶

• Teachers need to frame AI use around the distribution of epistemic responsibility rather than treating productive AI use as a matter of prompt formulation alone.

Constraints¶

• The article presents these as possibilities for making the consequences of different distributions of epistemic responsibility visible, not as tested interventions.

Target Learners¶

• secondary-school mathematics students

Target Learning Goals¶

• self-regulated AI use and independent mathematical-modelling performance

Affordances¶

• Epistemic Proactivity Framework

Related Strategies¶

• Scenario A: designing and deconstructing AI prompts as an epistemic practice

Examples¶

-

Key Sources¶

• Abdelghani, R., Kaiser, P., Murayama, K. (2026). Changes in Help-Seeking Strategies Predict unaided Performance during AI-based Mathematical Learning. https://arxiv.org/abs/2606.28472. doi:10.48550/arxiv.2606.28472