Artificial intelligence as a pedagogical communicator: mixed-methods insights from Raffles International College Bangkok

Artificial intelligence (AI) is increasingly employed as a pedagogical assistant in higher education, but its role as a communicator in diverse classrooms is not fully understood. This mixed-methods study investigates AI as a pedagogical communicator at a private international college in Bangkok. The study extends the Technology Acceptance Model by integrating social presence and trust constructs to examine factors influencing student engagement and satisfaction with AI-mediated instruction. In an explanatory sequential design, a survey of 300 students was analyzed using structural equation modeling, followed by 10 in-depth interviews. Results indicated that student satisfaction is driven by two mediated pathways: perceived social presence in AI communication builds trust, which enhances satisfaction, and perceived usefulness of AI feedback promotes engagement, which likewise increases satisfaction. Ease of interacting with the AI increased perceived presence and usefulness and thus indirectly boosted satisfaction; however, it also raised expectations that sometimes dampened satisfaction. Interview themes clarified these patterns. Students valued the AI’s clear explanations, responsiveness, and round-the-clock support, which improved efficiency and engagement, but also noted its lack of human warmth and increased pressure from constant availability. These findings suggest design principles for AI communicators, such as using friendly, context-aware language and clear rationales to build trust, and providing stepwise, practical feedback to sustain engagement. In sum, culturally responsive AI tutors that balance human-like connection with instructional efficiency can enhance student engagement and satisfaction in multicultural educational settings.