Purpose: The rapid integration of Artificial Intelligence (AI) into mathematics education risks fostering "cognitive offloading," in which students passively rely on automated answers rather than developing critical reasoning. Despite this growing concern, prior studies have largely ignored how combining AI with structured, high-impact pedagogy might prevent this technological dependency. Addressing this gap, this study evaluates how integrating GASING pedagogy with AI tools enhances problem-solving proficiency. Specifically, it simplifies complex learning models to investigate how student motivation and mathematical creativity actively mediate the relationship between pedagogy, technology, and cognitive outcomes. Method: Employing an explanatory sequential mixed-methods design, quantitative data were collected from 120 secondary school students (Grade 8, ages 13-14) in Jakarta, Indonesia. The structural relationships were tested using Partial Least Squares Structural Equation Modeling (PLS-SEM) via SmartPLS 4. This was followed by a focused qualitative thematic analysis of semi-structured interviews and observations to explain the mechanisms driving the statistical results. Findings: The results confirm a serial mediation model in which GASING Pedagogy positively influences Learning Motivation, which, in turn, enhances Mathematical Creativity. Furthermore, Mathematical Creativity was found to be a significant mediator between AI Adoption and Problem-Solving proficiency. Qualitatively, students perceive the synergy between GASING and AI as simplifying complex tasks and stimulating divergent thinking, positioning human creativity as a critical 'gatekeeper' for validating AI-generated outputs. Significance: This research shifts the discourse from AI dependency toward a distributed cognition framework. Theoretically, it establishes creativity as an essential mediator in digital mathematics pedagogy. In practice, it provides educators with a protocol for balancing automated tools with pedagogical scaffolding, cultivating adaptive problem-solvers capable of rigorous intellectual verification.
Copyrights © 2026