COMPUTERS ARE SOMETIMES DISTINCT AND SOCIAL ACTORS: RESPONSES TO OPINION CONFORMITY FROM HUMANS AND COMPUTERS

Open Access
- Author:
- Gambino, Andrew
- Graduate Program:
- Mass Communications
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- October 08, 2021
- Committee Members:
- Matt McAllister, Major Field Member
Denise Solomon, Outside Unit & Field Member
Anthony Olorunnisola, Program Head/Chair
Michael Schmierbach, Chair & Dissertation Advisor
Mary Oliver, Major Field Member - Keywords:
- HCI
CASA
Communication
Communication and Technology
Social Presence
Chatbot - Abstract:
- Abstract The increased use of increasingly social technologies requires investigation. The computers are social actors paradigm (CASA; Nass & Moon, 2000) has frequently been used to guide research of digital social technologies. The CASA proposition that we treat computers like people has been confirmed through empirical studies in modern contexts such as chatbots (Ho et al., 2018) and social robots (Lee & Liang, 2019). Despite successful application into a wide range of contexts, there has also been considerable research that suggests a need to reconsider CASA and its assumptions (Gambino et al., 2020). In this dissertation, a test of CASA and Gambino et al’s (2020) extended CASA are tested in the context of a live customer service chat. Opinion conformity was selected as a social cue due to its documented persuasive interpersonal effects as well as its social inappropriateness from a computer source. According to CASA, opinion conformity was predicted to have a positive effect on persuasive outcomes for both human and computer sources. Alternatively, based on Gambino et al. (2020), opinion conformity was predicted to have a positive effect from humans but a negative effect from computers. The results of the two-factor experiment (source: human, computer; opinion conformity: absent, present) did not suggest support for either perspective. Contrary to predictions, participants responded negatively to opinion conformity from human sources. Additionally, participants responded positively to opinion conformity from computers. This pattern was moderated by participant preference for virtual assistants, with the two-way interaction effect dissipating at higher levels, which provides some support for the extended CASA. Incorporating prior use and experiences is recommended for future research in the CASA domain. As interactions with and through social technologies continue to increase, the study of technologies can provide insight into our interpersonal expectations and relationships. To accomplish this, it is first necessary to incorporate our use of social technologies into our communication and technology theories.