To augment or to automate: impact of anthropomorphism on users’ choice of decision delegation to AI-powered agents
Shaista Anayat, Arun Kumar Kaushik
Abstract
Understanding how anthropomorphic design influences user decision-making is critical as anthropomorphic technologies become more integrated into daily life and work environments. The existing literature on AI-anthropomorphism mainly focuses on physical or voice-based anthropomorphism and its impact on adoption. However, it ignores mind-based (cognitive and affective) anthropomorphism and its impact on users’ decision delegation (automated or augmented) to AI-systems. This study examines how mind-based anthropomorphism influences users’ decisions to delegate tasks to AI-powered agents, using a survey of 243 actual AI agent users and PLS-SEM for data analysis. Results revealed that users’ choice for augmented decision delegation to AI-powered agents is significantly increased by cognitive anthropomorphism. Also, users’ choice for automated decision delegation to AI-powered agents is positively influenced by affective anthropomorphism. Trust emerged as a crucial mediator between anthropomorphism and automated decision delegation, encouraging users to hand over the decision-making completely to AI-powered agents. Results also revealed decreased reliance on trust when users prefer to retain a certain level of control in the decision-making process. The study significantly contributes to the literature on human-AI interaction by enhancing the understanding of users’ psychology towards anthropomorphism and their choice for decision delegation to AI-powered agents.