A Taxonomy of Social Cues for
When using the taxonomy, please cite as Feine, J., Gnewuch U., Morana S. and Maedche A. (2019): “A Taxonomy of Social Cues for Conversational Agents” International Journal of Human-Computer Studies. To read the paper, please click here.
A facial expression consists of one or more motions/movements on the face of the agent.
Smiling, looking angry, moving eyebrow.
Agents can express their emotions using facial expression (Becker et al. 2004; Becker et al. 2005;Guo et al. 2016) and are preferred over nonrealistic static agents (McBreen et al. 2001). Also, facial expressions impact the credibility of agents and outweigh other body cues (Cowell, Stanney 2005). Mirroring of the user’s facial expression can impact the acceptance, trust and likeability of agent (Lisetti et al. 2013). Also, expressed facial emotions of an agent are perceived different dependent on the age of the user (Beer et al. 2015). Furthermore, facial expression can indicate turn-taking attempts (Cassel, Bickmore 2000), and positive or negative facial expression create similar impression on anxiety like in real world (Pertaub et al. 2001; Gebhard et al. 2014). Reactive laughing leads to an emotional contagion effect (Niewiadomski et al. 2013), increases social presence (Pecune et al. 2015) and is perceived as more positive than smiling agents (Ding et al. 2014). Smiling agents impacts perceived likability (Nunamaker et al. 2001; Cafaro et al. 2016), let users smile longer when interacting with the agent (Kramer et al. 2013), impact the performance on a task (Kraemer et al. 2016), and is useful as part of relational behavior strategies to ensure a long-term working alliance (Bickmore et al. 2005; Bickmore et al. 2010). Facial expression can also distract the user leading to decreased user involvement (Hess et al. 2005). Also, people respond to emotional facial expression different dependent on the agent’s gender (Hayashi 2016). Specific facial expressions can further be used for flirtation which impacts the user’s enjoyment, increased their interest to continue the interaction or even to engage in a conversation (Bee et al. 2009).