Someone to watch over me: presence of an assistant agent on SNS inhibits negative blaming statements in tweeting

This study explored whether artificial cues of others on social networking sites (SNS) have an effect on a user's tweeting behavior in frustrating situations. By manipulating artificial cues of others, we measured the statements posted by participants on a hypothetical SNS under two different experimental situations: one was the agent condition, in which task instructions were given by an assistant agent throughout the task, and the other was the control condition in which task instructions were given through a text box. In each condition, participants were first asked to read fictional stories including frustrating interpersonal and intrapersonal events and to adopt these situations as their own. Participants were then asked to tweet about their feelings on the hypothetical SNS. The negativity and aggressiveness of each statement in both conditions were assessed by other participants. Results demonstrate that the statements for the frustrating interpersonal event posted by participants under the agent condition were more often evaluated as non-negative and less often evaluated as negative-and-blaming-of-others than those posted under the control condition. Conversely, there was no difference between conditions in the statements for the frustrating intrapersonal event. These findings suggest that the implication of the presence of others, such as an assistant agent, in the tweeting environment inhibits disparaging tweeting behavior on SNS.