Trajectory-sensitive scaffolding: use recent interaction history to return epistemic work to the learner when closure-oriented requests accumulate¶

Strategy · All strategies
Evidence · no claims cited

Description¶

The authors propose that an AI tutor consider recent interaction history when deciding whether to intervene: "repeated or increasingly frequent requests for the AI to provide or confirm outcomes could motivate a scaffold that returns some of the epistemic work to the learner". Such scaffolds would ask students to articulate their current reasoning, identify what remains unclear, or attempt the next step before receiving further assistance, while distinguishing occasional, purposeful offloading from a developing pattern of delegation.

Design Implications¶

Context¶

Requirements¶

• The tutor must be able to detect changes in the orientation of requests across the interaction, not only isolated prompts or whole-session statistics.

Constraints¶

• The authors state this trajectory-sensitive scaffolding remains a design hypothesis to be experimentally tested rather than an intervention demonstrated by the present study.

Target Learners¶

• secondary-school students using general-purpose LLMs for learning

Target Learning Goals¶

• retaining epistemic responsibility and independent performance during AI-supported learning

Affordances¶

• Epistemic Proactivity Framework

Related Strategies¶

• Detect low-literacy Generation-Factual loops and intervene with prompts steering students toward drafting and requesting critique

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