When Being Human Backfires: Perceptions of Privacy Violations by Human vs Non-Human Service Agents
(A2025-126471)

AUTHORS

Aleksandra Mikhailova, Stockholm School of Economics; Magnus Söderlund, Stockholm School of Economics

ABSTRACT

In the age of rapid technological advancement, non-human service agents are becoming integral to consumers' lives, raising critical questions about privacy and trust. Across two experimental studies (N = 378), findings reveal that human agents are perceived as more privacy-invasive than non-human agents and lead to greater trust erosion in service providers. A serial mediation analysis demonstrates that these effects are driven by mind attribution and perceived privacy violation severity, revealing a paradox: agents attributed with more mind initially foster trust but also amplify sensitivity to privacy violations. Despite the significant data-handling and analysis capabilities of non- human agents, results indicate a relative bias favoring machines, as they are perceived as less harmful. This study offers novel theoretical insights into privacy dynamics, practical implications for protecting consumer trust, and emphasizes the ethical imperative for proactive privacy safeguards as non-human agents continue to proliferate.