ChatGPT Overreliance, Writing Skill Development, and Academic Integrity: Evidence from Senior High School Learners

Abstract

This study examined the effects of ChatGPT overreliance on the writing skills and ethical perceptions of senior high school learners at Sta. Lucia National High School Dolores Quezon amid the growing integration of AI tools specifically ChatGPT in academic writing. Using a mixed-methods design, the research employed a descriptive survey (n = 102), three writing performance tasks analyzed through one-way MANOVA, and five paired t-tests assessing changes following an AI Awareness Intervention, supplemented by thematic analysis of student and teacher interviews. Quantitative results showed significant differences across high-, moderate-, and low-reliance groups in key writing competencies, including reading-to-writing transfer, thesis development, and argumentation, while the intervention produced significant improvements in ethical awareness and responsible use. Qualitative findings revealed that students use ChatGPT for convenience, stress reduction, and language support, but excessive reliance reduced cognitive engagement, limited idea generation, and raised concerns about plagiarism and authorship. Overall, the study found that although ChatGPT can scaffold writing processes, overreliance undermines deeper learning and academic integrity, whereas structured AI literacy interventions effectively strengthen ethical judgment and self-regulation. These results highlight the need for institutional policies and pedagogical frameworks that promote responsible and educationally meaningful integration of AI in writing instruction.