Alex Liu, Min Sun, Lief Esbenshade, Victor Tian, Zachary Zhang, Kevin He (2026) — arXiv. 📄 Full text (arXiv)
This large-scale K-12 deployment provides empirical evidence that teacher-authored prompts can reliably shape the cognitive quality of student-AI dialogue at classroom scale. The TASD system lets teachers define both the AI's role and the student-facing conversation starter, creating a two-layer orchestration model that produced 71% alignment with instructional goals. The 38% under-reach rate in cognitive demand—approaching 50% for the highest DOK level—mirrors known gaps between intended and enacted scaffolding documented in scaffolding and student-ai-interaction research. The intervention study showed that adding explicit finish lines to prompts and 'no direct answers' guardrails meaningfully narrowed the gap, connecting prompt engineering to formative-assessment design principles. The work also extends teacher-role scholarship by showing that teachers can operate as prompt architects without deep technical expertise, provided the system surfaces the right levers. This positions teacher-authored configuration as a bridge between generative-ai capabilities and k-12 classroom constraints.
Related Pages
- automated-grading — AI-generated feedback as scaffolded output
- ai-feedback-quality — Quality benchmarks for AI tutoring feedback
- teacher-role — Teacher orchestration of AI classroom tools
- student-ai-interaction — Scaling classroom student-AI dialogue
- higher-ed — Higher education feedback provision context
- feedback-loop — Discretionary feedback as high-value loop