Are emojis better? The impact of generative AI emoji cues and service outcomes on user satisfaction: evidence from ERPs

Abstract

To promote the sustainable development of Generative Artificial Intelligence (GenAI) applications in the service industry, enhancing user satisfaction is key. Emojis serve as catalysts for conveying emotions and enhancing user experience in online communication. However, due to the black-box nature and unpredictability of GenAI, service providers find it challenging to control the boundaries of their application. Currently, there is ongoing debate within the academic community regarding the use of emojis in GenAI, particularly concerning emotional manipulation and experience enhancement, with further exploration needed into their effectiveness and underlying mechanisms. This study is based on the emotion as social information model and employs event-related potential (ERP) technology with high temporal resolution, which is more suitable for GenAI interaction scenarios. By measuring users’ immediate cognitive processing and psychological activities, the study analyzes the underlying cognitive neural mechanisms through which emojis (presence vs. absence) and service outcomes (success vs. failure) influence user satisfaction. Behavioral results indicate that the outcome of GenAI services determines user satisfaction, while the presence or absence of emojis does not directly impact satisfaction. ERP results show that the presence of emojis compared to their absence triggers larger P3 amplitudes (emotional arousal) and N4 amplitudes (cognitive conflict); compared to service success, the presence of emojis during service failure triggers larger N4 amplitudes. This study reveals the complexity of user responses in real human-machine interaction environments, enriches research on the use of Emojis in GenAI, and provides scientific theoretical and practical foundations for GenAI design and enhancing user experience.