Abstract
Anthropomorphism is increasingly integrated into in-vehicle virtual agents (IVAs) to enhance driver trust, engagement, and interaction in automated driving systems. However, discrepancies in conceptualization, implementation, and assessment methodologies hinder a comprehensive understanding of its impact. This systematic review, conducted using the PRISMA 2009 framework, synthesizes existing 13 empirical research to examine the definitions, applications, and effects of anthropomorphic IVAs across varying driving environments. Findings indicate that IVAs exhibiting humanlike attributes, including speech, facial expressions, and emotional responsiveness, facilitate user engagement and trust, particularly in highly automated contexts. However, excessive anthropomorphism may induce overreliance, thereby compromising driver vigilance in semi-automated settings. The study underscores the necessity for standardized evaluation frameworks to enhance methodological consistency and cross-study comparability. Future research should prioritize refining anthropomorphic IVA design to balance humanlike interaction with system transparency, ensuring both enhanced user experience and operational safety in evolving automated vehicle technologies.
Get full access to this article
View all access options for this article.