The Trusted Caregiver: The Influence of Eye and Mouth Design Incorporating the Baby Schema Effect in Virtual Humanoid Agents on Older Adults Users' Perception of Trustworthiness

TL;DR

This study investigates how eye and mouth design, inspired by the baby schema effect, influences older adults' perception of trustworthiness in virtual humanoid agents, providing a design paradigm to enhance trust in smart home care.

Contribution

It introduces a novel design framework for facial features of virtual agents based on the baby schema effect to improve trustworthiness among older users.

Findings

Aligning facial features with the baby schema effect increases perceived trustworthiness.

Specific proportions of eye and mouth sizes significantly enhance credibility.

The proposed design paradigm can be applied to develop more trustworthy virtual agents.

Abstract

The increasing proportion of the older adult population has made the smart home care industry one of the critical markets for virtual human-like agents. It is crucial to effectively promote a trustworthy human-computer partnership with older adults, enhancing service acceptance and effectiveness. However, few studies have focused on the facial features of the agents themselves, where the "baby schema" effect plays a vital role in enhancing trustworthiness. The eyes and mouth, in particular, attract most of the audience's attention and are especially significant. This study explores the impact of eye and mouth design on users' perception of trustworthiness. Specifically, a virtual humanoid agents model was developed, and based on this, 729 virtual facial images of children were designed. Participants (N=162) were asked to evaluate the impact of variations in the size and positioning of the eyes and mouth regions on the perceived credibility of these virtual agents. The results revealed that when the facial aspect ratio (width and height denoted as W and H, respectively) aligned with the "baby schema" effect (eye size at 0.25W, mouth size at 0.27W, eye height at 0.64H, eye distance at 0.43W, mouth height at 0.74H, and smile arc at 0.043H), the virtual agents achieved the highest facial credibility. This study proposes a design paradigm for the main facial features of virtual humanoid agents, which can increase the trust of older adults during interactions and significantly contribute to the research on the trustworthiness of virtual humanoid agents.