AI in School EFL Learning: A Systematic Review of Impact Pathways for Engagement, Achievement, and Satisfaction

Background: Artificial Intelligence (AI) is increasingly integrated into school-based English as a Foreign Language (EFL) instruction, yet the mechanisms through which it shapes learners’ engagement, achievement, and satisfaction remain insufficiently theorised. Although prior studies report positive effects, they rarely explain how AI influences learning processes within authentic classroom conditions. Purpose: This systematic review synthesizes empirical evidence to explain how and under what conditions AI technologies shape engagement, achievement, and satisfaction in school EFL contexts. Specifically, it aims to identify the mediating cognitive, affective, and behavioural mechanisms through which AI operates, examine contextual moderators influencing its effectiveness, and develop an integrative AI Impact Pathways Framework to guide theory-driven research and context-sensitive pedagogical design. Method: A comprehensive search across seven databases yielded 99 records; following PRISMA 2020 procedures, 23 empirical studies involving direct AI use by K–12 EFL learners were retained. Thematic synthesis was employed to identify cross-study patterns and inductively develop a multi-pathway explanatory framework. Findings: AI tools—including NLP-based feedback systems, intelligent tutoring systems, conversational agents, gamified applications, and adaptive learning platforms—enhanced engagement by increasing interactivity, reducing anxiety, and sustaining time-on-task. Achievement gains were associated with personalised scaffolding, iterative feedback loops, and opportunities for authentic language use across speaking, reading, writing, and vocabulary learning. Satisfaction improved when AI supported autonomy, emotional reassurance, and perceptions of usefulness. Three interrelated pathways—cognitive, affective, and behavioural—mediated these outcomes, while teacher readiness, digital infrastructure, cultural–linguistic fit, student digital literacy, and cognitive load served as five significant contextual moderators. Implications: The review advances theoretical understanding by proposing the AI Impact Pathways Framework in School EFL Learning, which clarifies how AI affordances interact with pedagogical processes and contextual conditions to shape learner outcomes. The findings provide guidance for designing equitable, context-sensitive AI integration in schools and highlight the need for longitudinal, cross-cultural, and theory-driven research.