Digital Matchmakers: How Chatbots Use Positive Emotions to\nInfluence Human Self-Disclosure

Despite the growing popularity of chatbots, their success depends on humans disclosing personal and sensitive information. Prior research indicates that human-like chatbots enhance self-disclosure by subconsciously applying social scripts and mirroring user characteristics. However, the role of emotions in chatbot-human interactions, especially regarding privacy tradeoffs, remains complex. This study aims to improve understanding of how chatbots’ use of positive emojis influences self-disclosure in online dating contexts. We developed an experimental matchmaking chatbot, TinderBot, and conducted two randomized experiments with 759 participants. Experiment 1 manipulated the chatbot’s use of emojis, while experiment 2 also varied the sensitivity of personal questions. The results suggest that chatbots using positive emojis significantly increase perceived friendliness of the chatbot, which positively influences users’ emotional state and self-disclosure behavior. This effect is substantial for highly sensitive questions but negligible for less sensitive ones. Our findings highlight the potential of emotion-driven chatbot design to foster trust and enhance user engagement and information sharing in digital matchmaking platforms, while also underscoring the importance of being aware of privacy tradeoffs.