Beyond the Prompt: Student Voice, Writer Identity, and Authorship Confidence in AI-Assisted Academic Writing

This study examined how artificial intelligence writing tools relate to senior high school students’ writer voice and identity confidence in academic writing in Tagbina 2 District, Surigao del Sur, Philippines. Using a convergent parallel mixed-methods design, the study collected quantitative data from 154 students and 15 teachers and qualitative data through focus group discussions with 10 students and semi-structured interviews with five teachers. Findings showed that students used AI writing tools at a moderate and purpose-driven level, primarily for grammar support, organization, and summarizing rather than full-text generation. Writer voice remained generally evident in AI-assisted writing, with style and vocabulary more strongly preserved than authenticity and tone. Students reported generally positive identity confidence, especially in ethical clarity and writing self-efficacy, while ownership confidence emerged as the weakest dimension. Teachers’ perceptions were broadly consistent with students’ self-reports, and no significant differences were found between the two groups in their evaluations of writer voice and identity confidence. Statistical analysis further showed that AI tool usage was significantly related to vocabulary, style, writing self-efficacy, and ethical clarity, but not to authenticity, ownership confidence, or revision agency. Qualitative findings portrayed AI as both a scaffold and a possible eraser of voice, highlighting authorship as a spectrum shaped by overreliance, academic pressure, unclear guidelines, and teacher mediation. Overall, the study suggests that AI can support writing development when used critically and ethically, but preserving authentic voice and authorship requires student agency, sustained teacher guidance, and clear institutional direction.