Zhengxu Li, Junling Wang, April Yi Wang (2026) — ACM L@S 2026. Study of when teachers should control AI-generated classroom visuals.
Key Findings
Generative AI can help teachers rapidly create classroom-ready visual materials, particularly in mathematics where diagrams and visual representations must be pedagogically meaningful and instructionally correct. This paper investigates when and how teachers should control AI generation of mathematical visuals.^[raw/papers/2605.10672.md]
The study identifies key decision points where teacher input is essential: ensuring pedagogical correctness of visual representations, alignment with curriculum goals, and appropriateness for student levels. The authors propose a framework for teacher-AI collaboration in visual material creation that balances AI efficiency with pedagogical control.^[raw/papers/2605.10672.md]
Connections to AIED
This work connects to teacher-role by showing how teachers remain essential as pedagogical validators of AI-generated content. It intersects with stem-education since mathematics visual generation is a core need in STEM teaching.^[raw/papers/2605.10672.md]
The findings also relate to ai-generated-content and principled-ai-education, suggesting that principled AI use in education requires human oversight for pedagogically sensitive outputs. The teacher control framework connects to agentic-workflows-education where AI agents assist but humans direct.
Related Pages
- teacher-role — Teachers as pedagogical validators of AI-generated math visuals
- stem-education — Mathematics visual generation as a core STEM teaching need
- agentic-workflows-education — Teacher-AI collaboration framework for visual material creation
- ai-generated-content — Pedagogical requirements for AI-generated educational content
- principled-ai-education — Human oversight as a principle for AI use in education
Citation
APA: Li, Z., Wang, J., & Wang, A. Y. (2026). When should teachers control AI generation for mathematics visuals? arXiv:2605.10672. Proceedings of the Thirteenth ACM Conference on Learning @ Scale (L@S 2026), Seoul, Republic of Korea.