TY - GEN
T1 - GenAI-Supported Mathematical Storytelling
T2 - 27th International Conference on Artificial Intelligence in Education, AIED 2026
AU - Qiu, Jingyuan
AU - Chang, Sheng
AU - Wei, Wei
AU - Chen, Ziqi
N1 - Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2027.
PY - 2027
Y1 - 2027
N2 - Generative AI (GenAI) has the potential to enable efficient creation of personalized mathematical storytelling, yet its optimal depth of integration across cognitive levels remains unclear. This study investigated how three GenAI-supported mathematical storytelling integration levels (Low-cognitive, High-cognitive, and Full Integration) impacted secondary students’ cognitive engagement and knowledge acquisition. A mathematics teacher utilized a multi-agent system to co-create level-specific geometry narratives contextualizing the reflection principle. Analyzing survey and post-test data from 112 Grade 8 students via ANCOVA and MANCOVA, results revealed that the LI condition fostered significantly higher cognitive engagement than the HI and FI groups. Furthermore, significant knowledge differences emerged only at the foundational “knowing” level, where both LI and FI groups outperformed the HI group. These findings indicate that integrating GenAI at lower cognitive levels is optimal, providing evidence-based practical guidance and a theoretical framework for deploying level-specific, GenAI-assisted mathematical storytelling.
AB - Generative AI (GenAI) has the potential to enable efficient creation of personalized mathematical storytelling, yet its optimal depth of integration across cognitive levels remains unclear. This study investigated how three GenAI-supported mathematical storytelling integration levels (Low-cognitive, High-cognitive, and Full Integration) impacted secondary students’ cognitive engagement and knowledge acquisition. A mathematics teacher utilized a multi-agent system to co-create level-specific geometry narratives contextualizing the reflection principle. Analyzing survey and post-test data from 112 Grade 8 students via ANCOVA and MANCOVA, results revealed that the LI condition fostered significantly higher cognitive engagement than the HI and FI groups. Furthermore, significant knowledge differences emerged only at the foundational “knowing” level, where both LI and FI groups outperformed the HI group. These findings indicate that integrating GenAI at lower cognitive levels is optimal, providing evidence-based practical guidance and a theoretical framework for deploying level-specific, GenAI-assisted mathematical storytelling.
KW - cognitive engagement
KW - cognitive level
KW - generative AI
KW - lesson plan
KW - mathematics storytelling
UR - https://www.scopus.com/pages/publications/105043966884
U2 - 10.1007/978-3-032-29770-9_2
DO - 10.1007/978-3-032-29770-9_2
M3 - Conference contribution
AN - SCOPUS:105043966884
SN - 9783032297693
T3 - Lecture Notes in Computer Science
SP - 11
EP - 19
BT - Artificial Intelligence in Education - 27th International Conference, AIED 2026, Proceedings
A2 - Blanchard, Emmanuel G.
A2 - Chen, Guanliang
A2 - Chi, Min
A2 - Isotani, Seiji
PB - Springer Science and Business Media Deutschland GmbH
Y2 - 27 June 2026 through 3 July 2026
ER -