Effectiveness of AI-Supported Learning in English Vocabulary Acquisition among Undergraduate Learners: A Quasi-Experimental Study

Abstract This study examines the effectiveness of structured AI-supported learning in English vocabulary acquisition among undergraduate learners. A quasi-experimental design involved two intact B.A. Part-I student divisions at Shri Vijaysinha Yadav College, Peth Vadgaon. Both groups received instruction covering 100 target vocabulary items in four areas. The Experimental Group received teacher-guided support through an AI-Supported Vocabulary Learning Cycle using NotebookLM and ChatGPT for learning, practice, feedback, and revision. Both groups subsequently completed the same 100 item vocabulary post-test. Independent samples t-tests, one-way ANOVA, and Cohen’s d were used. The Experimental Group achieved a significantly higher overall mean score (75.53) than the Control Group (66.95), t(70) = 6.39, p < .001, d = 1.51. All four categories showed significant differences, strongest for phrasal verbs. Gender showed no significant difference, whereas previous AI exposure and regular smartphone use were associated with higher achievement among Experimental Group learners. The findings suggest that AI can contribute to vocabulary learning when carefully and purposefully integrated with teacher-led classroom instruction and curriculum objectives. AI extends practice, feedback, revision, and application. Overall findings should be interpreted cautiously because of the intact-group design and absence of delayed assessment.