Lived Experiences of High School Students on the Use of Large Language Models Toward A Proposed Evidence-Based AI Implementation Framework

This qualitative phenomenological study explores the lived experiences of junior high school students at Pinamukan Integrated School, Batangas City, regarding their use of Large Language Models (LLMs) for academic tasks. Guided by Colaizzi’s method, in-depth interviews with twelve purposively selected participants revealed that students consistently integrate LLMs into their daily routines, valuing them for research, idea generation, and task efficiency. LLMs act as cognitive scaffolds, enhancing learning accessibility and reducing workload. However, students also recognize significant limitations, including inaccuracies, prompt sensitivity, and output inconsistency, necessitating critical evaluation. The use of LLMs presents dual emotional impacts, fostering confidence and creativity while simultaneously provoking anxiety, guilt, and concerns over intellectual deskilling. Students demonstrate a balanced approach, advocating for responsible use that complements, rather than replaces, personal effort and critical thinking. Based on these findings, the study proposes an Evidence-Based AI Implementation Framework designed to guide ethical, reflective, and educationally sound integration of LLMs in the classroom, aligning with DepEd’s mandate for responsible AI adoption in Philippine education.