Would you rely on an eerie agent? A systematic review of the impact of the uncanny valley effect on trust in human-agent interaction

Abstract

Trust is a fundamental component of human-agent interaction. As artificial agents become more common in daily life, it is essential to understand how people perceive and trust them. One key challenge is the Uncanny Valley Effect (UVE), in which increasingly humanlike agents can appear eerie or aversive. Despite growing interest in both trust and the UVE, research varies widely in how these concepts are defined and measured, making it unclear when and how the UVE affects trust. A systematic account of their relationship is therefore still missing. This review examines the impact of the UVE on human trust in artificial agents, broadly conceptualized as a willingness to rely on an agent based on beliefs in its competence, good intentions, and reliability, even in situations where such reliance involves uncertainty or risk. Moreover, it identifies methodological patterns, limitations, and gaps in the empirical literature. Following PRISMA guidelines, a systematic search identified 53 empirical studies that investigated UVE-related constructs alongside trust or trust-related outcomes. We analyzed these studies using a structured set of categories, including agent types, interaction forms, methodological approaches, measurement strategies, and key findings. The review shows that most studies rely on static images or hypothetical scenarios with limited real-time interaction, and that the majority use subjective trust measures vulnerable to bias. We further propose best-practice criteria for empirically investigating the UVE and its effects on trust. As the first systematic attempt to map this intersection, the review clarifies the current evidence base and provides a foundation for future research.