Abstract
This quasi-experimental study investigates the impact of Grammarly, an AI-powered writing assistant, on the confidence and self-efficacy of high school students in academic writing. Conducted with 48 students from a school of Physics and Mathematics, the research employed a non-randomized design with two groups: an Experimental Group using Grammarly Premium and a Control Group that did not receive technological support. Results indicated that Grammarly significantly reduced writing errors in the EG, compared to a lesser reduction in the CG. Additionally, the EG reported notable improvements in writing confidence and self-efficacy, with increases observed on validated scales. These findings suggest that Grammarly can substantially enhance both the quality of academic writing and students’ self-perceptions as writers. The study highlights the potential of integrating digital tools in educational settings to foster better learning outcomes. Limitations include the study’s short duration and non-randomized group assignment, pointing to the need for further research to explore long-term effects and broader educational impacts.