<![CDATA[The integration of Artificial Intelligence (AI) in mathematics education often treats it as a "black box" computational tool, neglecting students' understanding of inferential logic and uncertainty. This study addresses this gap in "AI Thinking" by formulating operational indicators for secondary mathematics, designing Bayesian Network (BN)-based learning scenarios for conditional probability, and evaluating BN’s role in supporting probabilistic reasoning. Using a Design-Based Research (DBR) approach over two iterative cycles with 64 eleventh-grade students, results indicate that BN effectively bridges abstract conditional probability with Directed Acyclic Graph (DAG) representations, enabling students to perform belief updates and "what-if" simulations, with significant post-test improvements (mean increase = 18.7; 0.84). The implications of this study provide a specific pedagogical contribution by shifting AI literacy from general computer science into the core mathematics curriculum. By utilizing BN, this research demonstrates how complex Bayesian inference can be made accessible to secondary students, transforming the teaching of "uncertainty" from static formula-plugging into dynamic, model-based reasoning. This framework offers educators a scalable way to integrate transparent AI logic directly into statistics and probability lessons, empowering students to critically evaluate model assumptions and evidence-based decisions in an increasingly automated world.]]>
Copyrights © 2026
