Exploring Virtual Agents for Augmented Reality

Abstract

Prior work has shown that embodiment can benefit virtual agents, such as increasing rapport and conveying nonverbal information. However, it is unclear if users prefer an embodied to a speech-only agent for augmented reality (AR) headsets that are designed to assist users in completing real-world tasks. We conducted a study to examine users' perceptions and behaviors when interacting with virtual agents in AR. We asked 24 adults to wear the Microsoft HoloLens and find objects in a hidden object game while interacting with an agent that would offer assistance. We presented participants with four different agents: voice-only, non-human, full-size embodied, and a miniature embodied agent. Overall, users preferred the miniature embodied agent due to the novelty of his size and reduced uncanniness as opposed to the larger agent. From our results, we draw conclusions about how agent representation matters and derive guidelines on designing agents for AR headsets.

Wow, You Are Terrible at This!

“…When comparing a teddy bear agent, a human-like agent, and an agent with a box as a head, Bailenson et al [4] showed that the teddy bear and human-like agents were rated as more likeable than agent with box-like features. In addition, for agents in augmented reality applications, Wang et al [43] showed how agents that were perceived to be more human were more likeable and personable than one that appeared more like a machine.…”

“…We created a łfind-the-hidden-object gamež in which users interact with a virtual agent that gives hints and provides feedback on their performance. The task was inspired by prior work [16,44], which asked people to find objects while engaging with an agent that knew where the objects were located. To understand how agent appearance and user cultural background affect the perception of the agent's feedback, we evaluated participants' emotional reactions.…”

“…Prior work has shown that both the appearance of an agent [4,43] and the level of anthropomorphism [7,8,11,34] affect users' perceptions. In our study, three different agents (Human, Mouse, and Robot) share a similar level of anthropomorphism (i.e., a humanlike body) but differ in the appearance of their head, hands, and feet.…”

“…These questions included ratings for if the agent was helpful, personal, trustworthy, appropriate, likeable, and if the user would be willing to interact with the agent again (continued use). These questions were adapted from prior work on embodied agents [3,43] and which we refer to as Agent Rating Questionnaire (ARQ) in the rest of this paper.…”

“…There is considerable research on the effects of agent appearance and agent perceptions in terms of likeability and uncanniness [8,34,43]. However, it is unclear how an agent's appearance affects how people interact with it: is feedback from a mouse or a robot taken less serious than feedback from a human agent?…”