Ứng dụng trí tuệ nhân tạo đa phương thức trong hoạt động đào tạo: Kinh nghiệm quốc tế và khuyến nghị

Tóm tắt

This article analyses international experience in applying multimodal artificial intelligence to training activities, focusing on technologies that combine text, audio, images, video, sensor data, and virtual or augmented reality environments. The study employs a structured literature review and qualitative content analysis of academic publications and policy documents published between 2018 and 2025. The findings indicate that multimodal AI is being adopted in four major ways: simulation-based experiential learning, multimodal learning analytics, personalised learning and intelligent feedback, and multimodal content creation and assessment. Implementation effectiveness depends on digital infrastructure, teachers’ pedagogical digital competence, data governance, and ecosystem collaboration. Drawing on international evidence, the article proposes a five-component framework for Viet Nam, including policy and governance, technological infrastructure, human capacity, curriculum and learning resources, and ecosystem partnerships. The paper concludes with policy implications for promoting effective, responsible, and sustainable adoption of multimodal AI in Vietnamese higher education.