Comparative Analysis of Self-Developed AI Virtual Instruction Models in University of Fujairah

Abstract

This study investigates the effects of learning through AI-generated virtual instructors while incorporating a unique element - the ability for students to choose their admired role model or celebrity as a personalized learning guide in terms of face and voice. We aim to assess the potential enhancements in learning outcomes and performance when employing various test scenarios: face and audio instruction of the admired role model, audio of the admired role model but different face; face of the admired model but audio different; normal (physical) instructor instruction. Our research maintains a crucial focus on preserving the student-teacher relationship by exploring ways in which AI can complement traditional instructional methods. The study also investigates the advancement of AI technology in assisting instructors and engaging students effectively. For a test sample of 105 students for a 5 minutes’ class, students felt it as an interesting concept, but with a small sample and time of class, it will not be correct to say this method alone will bring drastic learning improvement. In the first instance, students were curious and highly attentive, which is evident on the result that 76% of students answered the questions correctly compared to other groups. This system can be further improved by providing body/hand gestures with AI instructors, or allowing students to customize their own AI instructor.