Paper Number
ECIS2026-2323
Paper Type
CRP
Abstract
As artificial intelligence (AI) systems become increasingly human-like, anthropomorphism offers a useful lens for understanding how users develop trust in such systems. Many of these systems allow users to customise their anthropomorphic features to foster personal relevance and trust building. However, little is known about how customisation influences users’ perceptions of these systems. Drawing on anthropomorphism theory, we conducted a 2×2 between-subjects online experiment with 254 participants to examine how allowing users to customise a conversational agent (CA) for stress management influences their trusting beliefs. The results show that, contrary to theoretical expectations, customisation strengthens rather than reduces users’ tendency to anthropomorphise the CA. Further analyses reveal that customisation enhances emotional attachment, which, together with perceived humanness, acts as a key mechanism influencing trusting beliefs. These findings extend our understanding of AI anthropomorphism by highlighting the psychological implications of granting users agency in the customisation process.
Recommended Citation
Hu, Shanshan; Schmidt-Kraepelin, Manuel; Thiebes, Scott; Pietsch, Felix; Zhou, Zhongyun; and Sunyaev, Ali, "The Impact Of Customising Anthropomorphic Conversational Agents On Users’ Trusting Beliefs" (2026). ECIS 2026 Proceedings. 9.
https://aisel.aisnet.org/ecis2026/ai_anthro/ai_anthro/9
The Impact Of Customising Anthropomorphic Conversational Agents On Users’ Trusting Beliefs
As artificial intelligence (AI) systems become increasingly human-like, anthropomorphism offers a useful lens for understanding how users develop trust in such systems. Many of these systems allow users to customise their anthropomorphic features to foster personal relevance and trust building. However, little is known about how customisation influences users’ perceptions of these systems. Drawing on anthropomorphism theory, we conducted a 2×2 between-subjects online experiment with 254 participants to examine how allowing users to customise a conversational agent (CA) for stress management influences their trusting beliefs. The results show that, contrary to theoretical expectations, customisation strengthens rather than reduces users’ tendency to anthropomorphise the CA. Further analyses reveal that customisation enhances emotional attachment, which, together with perceived humanness, acts as a key mechanism influencing trusting beliefs. These findings extend our understanding of AI anthropomorphism by highlighting the psychological implications of granting users agency in the customisation process.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.