Abstract
With recent advances in artificial intelligence and the metaverse, virtual agents have become increasingly autonomous and accessible. Due to their growing technological capabilities, interaction with virtual agents gradually evolves from a traditional user-tool relationship to one resembling interpersonal delegation, where users entrust virtual agents to perform specific tasks independently on their behalf. Delegating to virtual agents is beneficial in numerous ways, especially regarding convenience and efficiency. Still, it poses problems and challenges that may drastically harm users in critical situations. This chapter explores the trust and delegation relationships between users and virtual agents, introducing a trust-based conceptual model to abstract and differentiate users’ delegation decisions based on three major dimensions covering the impact of rationality, affection, and technology. Practical guidance for virtual agent designs and potential applications of the model for metaverse development are also presented, followed by an outlook and an overview of future research opportunities.