AI-Enhanced Web-Based Personalized English Learning System

This study presents an artificial intelligence (AI) enhanced web-based personalized English learning system to improve motivation and knowledge construction in mixed-ability classrooms. The three-tier architecture combines a Flask-based learning analytics dashboard, a micro-service recommender containerized in Docker, and a real-time MongoDB log pipeline. A two-week quasi-experiment (N = 94) was conducted in a Saudi secondary school; students were randomly assigned to an artificial intelligence-enhanced web-based personalized English learning system (N = 47) or the school's standard Web Learning Management System (LMS (N = 47). Results revealed significant gains in learning motivation (η2 = .31) and vocabulary post-test scores (Cohen's d = 0.82) and a reduction in drop-off rate from 23% to 6%. Lag-sequential analysis showed that adaptive feedback elicited a 1.8× higher-order peer dialogue. Teachers reported that the web dashboard reduced preparation time by 30% and provided actionable insights for differentiated instruction.