From Screens to AI-Driven Human-Robot Interaction: Comparing Students’ Perceptions of the Robots Pepper and Furhat for ChatGPT-Based Exam Preparation
摘要
ChatGPT has been predominantly used on the screens of electronic devices. Our study goes beyond the screen as we apply humanoid robots as tangible interfaces for ChatGPT. We explore and compare students’ perceptions of the robots Pepper and Furhat used as embodied ChatGPT-connected interfaces providing students with practical advice for exam preparation related to focus and time management. Our study with 30 undergraduate students compared preferences and explored positive and negative aspects of engaging with the two robots used as exam guides. We used the Human Robot Interaction Experience Scale (HRIES) to compare the perceptions of both robots. The results show consistently higher ratings for Pepper compared to the Furhat robot. Specifically, students rated Pepper as significantly more friendly than Furhat, while Furhat was perceived as significantly more creepy compared to Pepper. Furthermore, statistically significant differences were observed in students’ preferences, trust, and willingness to interact with each robot. Students expressed a higher liking, trust, and intent to interact with the ChatGPT-connected Pepper robot for exam preparation. This study contributes valuable insights into the use of ChatGPT with tangible robotic interfaces. The findings emphasize the significance of not only functional capabilities but also perceived friendliness, trust, and appeal of educational robots which lend a “face” to conversational AI services like ChatGPT in educational settings and beyond.