Augmenting expressivity of artificial subtle expressions (ASEs)

Unfortunately, there is little hope that information-providing systems will ever be perfectly reliable. The results of some studies have indicated that imperfect systems can reduce the users' cognitive load in interacting with them by expressing their level of confidence to users. Artificial subtle expressions (ASEs), which are machine-like artificial sounds to express the confidence information to users added just after the system's suggestions, were keenly focused on because of their simplicity and efficiency. The purpose of the work reported here was to develop a preliminary design guideline for ASEs in order to determine the expandability of ASEs. We believe that augmenting the expressivity of ASEs would lead reducing the users' cognitive load for processing the information provided from the systems, and this would also lead augmenting users' various cognitive capacities. Our experimental results showed that ASEs with decreasing pitch conveyed a low confidence level to users. This result were used to formulate a concrete design guideline for ASEs.