Impact of AI-Integrated Peer Teaching on Performance and Engagement of Teacher Trainees: A Quasi-Experimental Study

Abstract

This study examined the impact of artificial intelligence (AI)-integrated peer teaching implemented through a combination of AI tools (ChatGPT, Google Gemini, Claude, Canva, and Gamma) on the academic performance, engagement, and perceptions of teacher trainees, and proposes a theoretical account of AI as an additional scaffolding actor that redistributes feedback, participation, and pedagogical support within peer teaching. Employing a quasi-experimental pretest-posttest design with intact groups, the research involved 36 student-teachers from the Bachelor of Education Special Education programmes, assigned to either a traditional peer teaching group or an AI-integrated peer teaching group. Data were collected through achievement tests, an engagement scale, an observation checklist, and a trainee perception survey. Statistical analysis revealed that both groups showed improvement in academic performance, but the AI-integrated group demonstrated significantly higher gains (mean post-test score: 10.72 Vs. 6.83, out of 50; mean difference = 3.89, 95% confidence interval [2.37, 5.41]; p < 0.001, Cohen’s d = 1.73, a large effect). AI-integrated peer teaching was also associated with higher trainee engagement, motivation, conceptual clarity, and personalized learning experiences, and qualitative observations further indicated more equitable participation and collaborative engagement in the AI-integrated group, although these qualitative and teaching-efficacy findings are exploratory given the study’s non-randomised, non-blinded design and should be interpreted as preliminary. The study concludes that AI-integrated peer teaching offers a promising approach to teacher education, and recommends institutional investment in digital infrastructure and capacity building alongside further, more tightly controlled and longitudinal research to test the proposed scaffolding mechanism directly.