Frankenstein's Monster in the Metaverse: User Interaction With Customized Virtual Agents

Area of Science:

• Human-Computer Interaction
• Virtual Reality
• Artificial Intelligence

Background:

• Virtual agents are increasingly human-like due to advancements in AI, computer graphics, and Extended Reality (XR).
• These agents have significant potential to assist users across various domains like customer service, healthcare, and education within the metaverse.
• Current technology allows for user customization of virtual agents, but research on its impact on user interaction is limited.

Purpose of the Study:

• To investigate user requirements for virtual agent customization.
• To evaluate the effects of virtual agent customization on user interaction and perception in Virtual Reality (VR).

Main Methods:

• A two-stage user study with 30 participants was conducted.
• Data collected included self-reported evaluations (e.g., motivation) and interaction metrics (e.g., duration, items tried).
• User perceptions of agent customization and subsequent human-agent interaction in VR were analyzed.

Main Results:

• Users desire full customization for agents in personal social circles but prefer limited options or pre-configured agents for general services.
• Attributes like gender, clothing, and hair were deemed more important for customization than facial features (skin, eye color).
• Customization of virtual agents significantly improved perceived social presence, rapport, and trust.

Conclusions:

• The level of desired virtual agent customization is context-dependent, varying between personal and service-oriented interactions.
• User preferences highlight specific attributes that are more impactful for virtual agent customization.
• Customizing virtual agents in VR positively influences user experience by enhancing social connection and trust.