Artificial Intelligence in the Teaching - Learning Process and Students' Grit and Performance in Mathematics

Abstract

This study explored the interrelationships among teachers’ integration of Artificial Intelligence (AI) tools in instruction, students’ utilization of AI tools, grit (students’ persistence in learning mathematics), and mathematics performance. Specifically, the research examined the extent of AI tool integration across teaching phases, students’ AI utilization across learning stages, levels of grit, and mathematics performance, as well as the predictive roles of AI usage and grit in student achievement. Anchored in contemporary learning theories and technology integration models, the study adopted a mixed-methods explanatory sequential design using purposive sampling, involving 240 Grade 8 students and five public high school mathematics teachers in Camarines Sur. Data were collected through surveys and performance tests and analyzed using ANOVA and regression techniques. Findings revealed that teachers integrated AI tools only to a limited extent across the pre-active, inter-active, and post-active phases of instruction. Likewise, students consistently utilized AI tools across the cognitive, associative, and autonomous stages of learning, yet without corresponding gains in higher-order mathematical performance. While students demonstrated a generally high level of grit—particularly in goal commitment—lower persistence was evident in overcoming setbacks and sustaining long-term interest. Mathematics performance remained at a fair level, with strengths in computational skills but notable weaknesses in conceptual understanding and problem-solving. These results expose a significant theoretical gap in prevailing assumptions within AI-enhanced learning and educational technology literature that increased access to or frequency of AI tool use directly translates into improved academic performance. Contrary to deterministic and instrumental views of educational technology, the findings suggest that AI integration functions more strongly as an affective and motivational mediator, influencing students’ grit rather than directly enhancing achievement outcomes. This reveals a conceptual tension between technology-centered performance models and learner-centered developmental theories, where affective traits such as grit play a pivotal yet under-theorized role in mediating learning outcomes. The study also identified that teachers’ AI integration was influenced by tool reliability, user-friendliness, and inclusivity, underscoring a gap between policy-driven AI adoption frameworks and context-sensitive classroom realities, particularly in public school settings. In response to these tensions, the study proposed the TEACH-AI Framework, which reconceptualizes AI integration as a strategic, inclusive, and pedagogically grounded process that aligns teaching phases, learning stages, and affective development. Theoretical implications highlight the need to extend existing AI-in-education models to incorporate grit as a mediating construct rather than treating technology use as a direct predictor of achievement. Practically, the study recommends strengthening teacher training on pedagogically aligned AI use, implementing growth mindset and resilience programs, and deliberately leveraging AI-supported environments to foster both conceptual understanding and sustained perseverance. The TEACH-AI Framework and LEARN-AI Framework is further recommended for implementation and testing across disciplines to examine its broader applicability and theoretical robustness.