Background: The rapid proliferation of Artificial Intelligence (AI) tools—particularly large language model-based applications—has created unprecedented opportunities for enhancing English as a Foreign Language (EFL) instruction. Despite growing enthusiasm, empirical evidence on their differential impact across key language competencies remains limited, particularly in South and Southeast Asian higher education contexts.

Objective: This study investigates the influence of AI-assisted language learning tools on EFL undergraduates' vocabulary acquisition, academic writing proficiency, and intrinsic motivation in a quasi-experimental university setting.

Method: A quasi-experimental pre-test/post-test design was adopted involving 120 undergraduate EFL students (60 experimental, 60 control) drawn from two intact sections of an English Communication Skills course. The experimental group used AI writing assistants and vocabulary applications (GPT-based and adaptive flashcard tools) over twelve weeks, while the control group followed traditional instruction. Data were collected through standardized vocabulary tests, analytical writing rubrics, and a validated motivation scale (MSLQ-adapted). SPSS v.26 was employed for independent samples t-tests and ANCOVA.

Key Results: The experimental group demonstrated significantly higher post-test scores in vocabulary breadth (p <.001, Cohen's d = 0.84) and academic writing quality (p <.01, d = 0.72). Intrinsic motivation showed a moderate but significant improvement (p <.05, d = 0.51). No significant gender-based differences were observed.

Conclusion: AI-assisted instruction meaningfully enhances core EFL competencies. Pedagogical integration frameworks and faculty digital literacy programs are recommended for sustainable implementation.