The current study investigated how artificial intelligence (AI) can be used to assist in teaching writing in the English Language Arts classroom at the secondary level. The study found the effects of an AI-based writing aid on students’ performance, engagement, and self-efficacy and to investigate the perceptions of the teachers and students concerning the opportunities and challenges. It was mixed method research with cluster randomized controlled trial involving 641 students in 26 classes in eight public secondary schools. The intervention was done in 12 weeks assigning participants randomly to AI-assisted conditions and business-as-usual conditions. Descriptive statistics and two-level ANCOVA models were used to analyze quantitative data, and thematic analysis was used to analyze qualitative data (teacher and student reflections). Results showed significantly small to moderate increases in standardized ELA and analytic writing and a large positive effect on engagement although no significant increases in self-efficacy. Subgroup analyses showed more positive results among lower-achieving, higher-SES, and non-multilingual students, which concern the issue of equity. According to the qualitative findings, there is importance of timely feedback, students agency that increased under pressure, and taught workload relief, but also a lack of self-efficacy improvement and access difficulties. AI can be used to improve learning if supported by pedagogy and equity-conscious design. It is recommended that AI feedback should be combined with teacher scaffolding and AI should be used to foster reflective student agency.
Soomro et al. (2025) studied this question.