The impact of emotional expression by artificial intelligence recommendation chatbots on perceived humanness and social interactivity

Artificial intelligence-powered chatbots capable of expressing emotions have gained significant popularity in the realm of customer service. Although previous studies have explored the impact of emotional expression in chatbots, there is a lack of understanding regarding the precise effects of different emotional cues. In this study, we drew upon social presence theory to investigate how different emotional cues conveyed by recommendation chatbots affect perceived humanness, social interactivity, and social presence. We conducted a series of scenario-based online experiments to shed light on these dynamics. We found that all three emotional cues (text, emoticons, and images) employed by chatbots can increase perceived humanness and social interactivity. Social presence appears to be an underlying mechanism for these positive relationships. We also observed two-way interactions for any pair of emotional cues and a three-way interaction for all three emotional cues. Ultimately, to elicit the most favorable customer perception, we propose that a mode of emotional expression using either text or emoticons alone is most appropriate. These findings deepen our understanding of the impact of emotional expressions in chatbots and offer novel insights into how to deploy chatbots in customer service.