Paper Type
Short
Paper Number
PACIS2026-1429
Description
This study investigates how the source of AI disclosure—system-led versus social-led—shapes users’ psychological mechanisms and behavioral responses in a post-disclosure context. Drawing on Expectancy Violation Theory and Attribution Theory, this study hypothesizes that social-led disclosure triggers stronger negative reactions by evoking an integrity-based attribution, leading to greater expectancy violation, trust erosion, and a sense of betrayal, ultimately resulting in higher levels of Negative Word-of-Mouth toward the platform. To test the proposed model, a scenario-based between-subjects experiment was conducted via the CloudResearch platform. Preliminary findings from 60 participants indicate that social-led disclosure elicits greater trust erosion and a sense of betrayal compared to system-led disclosure. By conceptualizing the disclosure source as a distinct theoretical construct, this research contributes to the AI disclosure literature and offers strategic insights for platform governance and the implementation of AI transparency mandates.
Recommended Citation
Choi, Sunyoung and Zo, Hangjung, "How the Source of AI Disclosure Shapes Psychologcial Mechanisms and User Responses" (2026). PACIS 2026 Proceedings. 6.
https://aisel.aisnet.org/pacis2026/ai_ethic/ai_ethic/6
How the Source of AI Disclosure Shapes Psychologcial Mechanisms and User Responses
This study investigates how the source of AI disclosure—system-led versus social-led—shapes users’ psychological mechanisms and behavioral responses in a post-disclosure context. Drawing on Expectancy Violation Theory and Attribution Theory, this study hypothesizes that social-led disclosure triggers stronger negative reactions by evoking an integrity-based attribution, leading to greater expectancy violation, trust erosion, and a sense of betrayal, ultimately resulting in higher levels of Negative Word-of-Mouth toward the platform. To test the proposed model, a scenario-based between-subjects experiment was conducted via the CloudResearch platform. Preliminary findings from 60 participants indicate that social-led disclosure elicits greater trust erosion and a sense of betrayal compared to system-led disclosure. By conceptualizing the disclosure source as a distinct theoretical construct, this research contributes to the AI disclosure literature and offers strategic insights for platform governance and the implementation of AI transparency mandates.
Comments
03-EthicsSocietalImpact