VietTESOL International Convention 2026

This study investigates the effects of integrating Problem-Based Learning and generative artificial intelligence on EFL speaking development through AI-assisted video creation. Conducted in a private university in Southern Vietnam, the study involved 40 undergraduate EFL students. Participants completed a project in which they created a 90-second English video using HeyGen based on a problem-oriented speaking task. The activity required students to analyze a situation, generate ideas, prepare a script, and present their responses through an AI-generated avatar. A mixed-methods design was employed. Quantitative data were collected through a post-intervention questionnaire examining speaking improvement, target language use, engagement, motivation, confidence, and perceptions of the AI tool. Qualitative data were gathered through open-ended responses and semi-structured interviews. Descriptive statistics indicated positive student perceptions across all dimensions. Participants reported improvements in fluency, idea organization, confidence, and use of lesson-related language structures. They also perceived HeyGen as an accessible tool that reduced speaking anxiety and increased motivation to practice English. Qualitative findings further revealed that the combination of problem-solving tasks and AI-assisted video production promoted creativity, learner autonomy, and active engagement in speaking activities. The findings suggest that integrating generative artificial intelligence within Problem-Based Learning can provide meaningful opportunities for EFL speaking practice and may serve as an effective pedagogical approach in technology-enhanced language classrooms.