Date created: | Last Updated:

Category: Project

Description: In this study, participants engage in face-to-face interactions with a virtual agent in Virtual Reality. Interaction is realized via speech and output of the agents is driven via a Large Language Model, presented lip-synced in Virtual Reality. There are three tasks (within-subject): Small Talk (participants ask questions), knowledge test (participants have to answer knowlegde question but can ask the virtual agent for help), convincing (participants ask virtual agents for preferences and then have to convince the agent of the opposite). In additon we manipulate personality of the agents by changing context prompts of the LLM (Between-subject design): One agent is described as extraverted and friendly and keen so socialize, the other agent is described as introverted and less social.