3rd TAGRA

The rapid integration of artificial intelligence (AI) into education presents a critical challenge: institutions lack scalable, theoretically grounded tools to assess students' underlying cognitive and epistemic dispositions toward knowledge and technology. Understanding these dispositions is essential for personalizing learning and identifying potentially maladaptive patterns. This study addresses this gap by developing and applying a machine learning-based framework to evaluate two novel constructs in higher education students: Epistemic Orientation Index (EOI) and Mental Polarity Index (MPI). The EOI measures a student's position on a continuum from knowledge replication (viewing knowledge as fixed) to knowledge generation (viewing knowledge as evolving). The MPI assesses cognitive degradation and polarity shifts. To operationalize these indices, a supervised machine learning model was trained on a purpose-built database of textual phrases, each independently rated by experts for its alignment with EOI and MPI constructs. This trained model was then applied to analyze student-generated text data. The key findings demonstrate that the model successfully classifies students' epistemic and mental polarity profiles, revealing significant variability across the student population. This approach enables the automated identification of students who may benefit from pedagogical interventions tailored to their specific orientation, particularly those exhibiting patterns associated with cognitive rigidity or negative polarity. The significance of this research lies in providing a scalable, automated assessment tool that bridges theoretical models of cognition with practical educational applications. It offers a foundation for developing personalized teaching methodologies and early identification of students who may be at risk of disengagement or cognitive stagnation, thereby contributing to more adaptive and responsive higher education systems.