The Impact of AI Anthropomorphism on Citizens' Willingness for Emergency Co‐Production: Evidence From a Survey Experiment

ABSTRACT Although AI has garnered considerable attention for optimizing the efficiency of emergency governance, traditional emergency AI primarily focuses on the implementation of technical functions and often engages in human–computer interaction merely as an instrumental tool. It fails to address citizens' psychological perception needs in emergency scenarios by simulating human‐like traits, which Alimits its effectiveness in motivating citizens' willingness for emergency co‐production. To fully unlock the value of AI in emergency cooperation, this study, centred on post‐disaster rescue scenarios, explores how AI anthropomorphism enhances citizens' willingness for emergency co‐production. Empirical analysis, conducted via three rounds of randomized survey experiments, reveals that a high degree of AI anthropomorphism can significantly boost citizens' willingness for emergency co‐production and this effect is transmitted through the dual pathways of competence trust and emotional trust. Furthermore, technology acceptance moderates the impact of AI anthropomorphism: in contexts where technology acceptance is high, the positive promotional effect of AI anthropomorphism becomes more pronounced. This study breaks free from the traditional instrumental understanding of AI in emergency co‐production, reconceptualizes AI as an active interactive agent, extends the Stereotype Content Model to human AI interaction in emergency settings and provides empirical evidence for the digital and intelligent transformation of emergency governance.