This dataset includes multimodal data collected from 26 higher education students (8 male, 18 female; 22 master’s and 4 doctoral students) who participated in a structured 30-minute session. The session comprised a 20-minute interactive discussion with Kai, an operational prototype focused on topics about AI ethics, preceded by a 5-minute orientation to the virtual environment and controls, and concluded with a 5-minute oral assessment. The study utilized a Quest3 headset for the interaction. The procedure began with each consenting participant completing a pre-questionnaire to collect demographic information, background data, and prior experiences with Generative AI (GenAI) and Extended Reality (XR) applications, including Mixed Reality (MR) and Virtual Reality (VR). The Partner Model Questionnaire (Doyle, 2022) was administered both before and after the interaction to assess participants’ perceptions of the GenAI agent as a speech interface. Participants were provided with a brief explanation of the task by the researcher before being guided to the designated setup area. In this area, participants were assisted in wearing the Quest3 headset and positioned in front of two TP-LINK Tapo C225 video cameras, which recorded the session from third-person perspectives capturing full-body movements. Additionally, the Quest3’s built-in first-person view camera recorded the interaction from the participant's perspective, creating a comprehensive multimodal dataset.

2024