Can we build a relationship with artificial intelligence (AI)?: the influence of AI on organization-public relationships
As the technology has developed, organizations have started to adopt artificial intelligence (AI) to communicate with their publics. Previous studies on organizational-public relationships had an assumption that all communications by an organization were created by humans. However, with the changes in the technology, public relations scholars and professionals have started to pay more attention to the potential of AI agents for online communication in terms of organization-public relationship. This dissertation explored the role of agent type and other communication attributes (i.e., human voice and immediacy of solution) in online organization communication. To achieve the purposes of the study, two experiment studies were conducted. Study 1 analyzed the effects of agent type and voice type on organization-public relationship outcomes and publics’ social media engagement intention in a social media setting with a 2 (type of agent: human agent vs. AI agent) x 2 (tone of voice: conversational human voice vs. organizational voice) between-subjects design experiment. A total of 363 participants were recruited for Study 1. Study 2 explored the impact of agent type and immediacy of solution to a question in an online live chat setting with a 2 (type of agent: human agent vs. AI agent) x 2 (immediacy of solution: immediate solution vs. delayed solution) between-subjects design experiment with 371 participants. The results showed that organizational messages communicated by a human agent generated significantly higher perceived control mutuality among participants in the social media setting than messages communicated by an AI agent. The use of conversational human voice resulted in significantly stronger relational outcomes (i.e., trust, satisfaction, commitment, and control mutuality). In the online chat setting, there were no significant differences between human and AI agents on relational outcomes or social media engagement intention among publics; however, immediacy of solution created higher levels of perceived trust, satisfaction, control mutuality, and social media engagement intention. The mediating role of relational outcomes and interaction effects of the variables in each setting are discussed. By addressing the roles of agent type and communication attributes, this dissertation advances the literature on public relations and human-computer communication.