Article Preview
TopIntroduction
Agents are systems that automatically process a given task with decision making, learning, and autonomy (Jennings & Wooldridge, 1995; Kiesler, Powers, Fussell, & Torrey, 2008; Maes, 1994). In today’s market for household and personal gadgets, intelligent agents with interfaces that permit user interaction are rapidly being commercialized. When using an interactive interface, consumers can imagine the personality of the agent or be emotionally affected through simulated conversation (Nass & Moon, 2000). Therefore, user experiences with an interactive, intelligent agent are likely to differ from user experiences with more traditional electronic devices.
Intelligent agents with dialogue characteristics feature an interactive user interface (UI) that permits naturalistic communication, allowing users to feel as if they were conversing with a real person. These dialogue characteristics are anthropomorphic. Anthropomorphism is the tendency to regard nonhuman agents as human beings and to attribute actual or imagined actions to objects (Epley, Waytz, & Cacioppo, 2007).
Previous studies on anthropomorphism have focused on visual characteristics such as the shape and gestures of a robot (Bartneck, Bleeker, Bun, Fens, & Riet, 2010; Darling, 2017; Epley et al., 2007; Eyssel & Ruiter, 2012; Kiesler et al., 2008; Salem, Eyssel, Rohlfing, Kopp, & Joublin, 2013; Tondu, 2012; Vallverdu, Nishida, Yoshisama, Moran, & Lazare, 2018; Waytz & Cacioppo, 2010). However, few studies have explored dialogue characteristics, the personifying feature of various intelligent agents currently being commercialized more rapidly than robots.
The aim of the current study was to investigate the effect of dialogue characteristics on user experiences with intelligent agents. Previous studies have addressed topics related to the technology of intelligent agents (Emary & Hassanien, 2011; Honarvar, 2013; Roger, King, Russ, Lambert, & Reese, 2001; Singh & Dave, 2013). However, the results of these studies did not address the relationship between user satisfaction and intelligent agent product design. Therefore, we proposed a solution to improve user satisfaction by investigating user experiences with intelligent agents. Our study also expands research on anthropomorphism; in the case of intelligent agents, the anthropomorphic element of intelligent agents is dialogue (i.e., audible), whereas the anthropomorphic elements of robots typically include appearance and gesture (i.e., visible). Finally, the current study suggests new directions for intelligent agent product design in order to improve user experiences with dialogue characteristics.
Using customer value theory as a framework, we explored user experiences with intelligent agents characterized not by human-like, physical characteristics (e.g., robots) but by human-like, conversational characteristics (e.g., voice-interactive smartphones). We categorized dialogue characteristics into three dimensions of value: dialogue support (functional value), playfulness (emotional value), and social presence (social value). These characteristics were the independent variables of the study. To collect data, we conducted surveys with people who used Siri, the intelligent agent on mobile devices designed and marketed by Apple.