Presentation
Exploring Dimensions of Human Likeness in Conversational AI: Implications for Human Identity Threat and Dehumanization
DescriptionConversational AI (CAI) has evolved into a non-human entity capable of thinking, feeling, and building relationships like humans. While attributing humanness to CAI—known as anthropomorphism—can foster trust and intimacy, it also raises concerns regarding threats to human identity and the dehumanization of humans. Therefore, identifying which humanlike traits in CAI contribute to these outcomes is crucial for ethical AI development.
This study aims to identify humanlike traits in CAI that influence human identity threat and dehumanization and to propose design guidelines to mitigate these effects. An online survey of 323 participants was conducted. Factor analysis revealed four key dimensions of CAI humanlikeness: Experience, Communication & Memory, Social Connection, and Agency. Structural equation modeling (PLS-SEM) showed that Experience heightened both identity threat and dehumanization, whereas Agency mitigated dehumanization. Social Connection generally reduced identity threat but amplified it when combined with high Experience. Furthermore, younger individuals reported greater identity threat and dehumanization, highlighting the importance of considering user age.
By revealing the psychological structure underlying users’ perceptions of CAI anthropomorphism, this study provides insights for designing CAI systems that balance the benefits of humanlike traits while preserving human distinctiveness.
This study aims to identify humanlike traits in CAI that influence human identity threat and dehumanization and to propose design guidelines to mitigate these effects. An online survey of 323 participants was conducted. Factor analysis revealed four key dimensions of CAI humanlikeness: Experience, Communication & Memory, Social Connection, and Agency. Structural equation modeling (PLS-SEM) showed that Experience heightened both identity threat and dehumanization, whereas Agency mitigated dehumanization. Social Connection generally reduced identity threat but amplified it when combined with high Experience. Furthermore, younger individuals reported greater identity threat and dehumanization, highlighting the importance of considering user age.
By revealing the psychological structure underlying users’ perceptions of CAI anthropomorphism, this study provides insights for designing CAI systems that balance the benefits of humanlike traits while preserving human distinctiveness.
Event Type
Lecture
TimeThursday, October 16th3pm - 3:20pm CDT
LocationGrand C/D South
Human AI Robot Teaming (AI)