Presentation
Inflection Matters: How Vocal Inflection in Decision Aids Affects Perceptions of Health Suggestions
DescriptionThis study investigates how vocal inflection in decision aids with humanlike voice affects trust in automation-assisted health scenarios. Using a 2 (inflection: upward/downward) x 2 (stated automation reliability: 95%/45%) factorial design, participants will interact with decision aids that offer audio-recorded health advice related to diabetes management. The advice presented by voice assistants will be legitimate in all trials, thus actual reliability will be 100%. Participants will rate their trust and likelihood to follow the advice, along with perceptions of vocal confidence and tone for each inflection type. It is hypothesized that downward inflection will be perceived as declarative and will lead to higher trust and greater willingness to follow advice than upward inflection, which may suggest uncertainty. Individual differences in locus of control related to one’s health and propensity to trust automation will also be assessed and are expected to moderate rated trust and expected reliance on the aids. This study will provide insight into the influence of inflection on subjective trust and perceptions of suggestions delivered by automation aids. Unlike traditional visual anthropomorphism, inflection may have utility in fostering positive human-automation interactions due to its potential to deliver information in a naturalistic manner, convincingly mimicking human attributes.
Event Type
Lecture
TimeThursday, October 16th4:50pm - 5:10pm CDT
LocationGrand Hall L
Aerospace Systems
Student Forum