VietTESOL International Convention 2026

The current research examines how EFL students engage with artificial intelligence (AI) within their Mobile English Learning Resource (MELR) ecologies for self-directed vocabulary learning (SDL). Using the Activity Theory (AT) framework, the research examines how AI influences students’ vocabulary learning processes and potentially alters their learning activity systems. A qualitative approach was employed in the study involving 22 Vietnamese EFL students, who kept learning journals for 7 weeks and participated in semi-structured interviews and focus group discussions for data collection. The data were then subjected to a thematic analysis. The research found that AI learning tools were most often integrated into MELR ecologies rather than used in isolation. Within the various resource ecologies, AI tools were used to help students construct, contextualize, explain, and automate strategies, generate practice, and organize learning activities. Also, AI integration shifted the burden of procedural and cognitive learning from the construction of learning resources to the management of learning activities. Nevertheless, several conflicting approaches persisted regarding the use of AI tools in lexical learning. The study also reveals several developmental pathways in students’ SDL of vocabulary, which illustrate students’ continual management of oppositions among learning objectives, self-regulation, contextual variations, and mediational tools. This study expands AT by illustrating how AI is reshaping mediation, redistributing the division of labor, and catalyzing transformations in self-directed vocabulary learning systems. It also provides new perspectives on AI-assisted vocabulary learning by positioning AI as an integrated, adaptable mediational tool within evolving MELR ecologies.