Technology for good? Rethinking human–AI synergy in normative messaging for sustainable consumption

Purpose This study aims to explore how human and robot service personnel can influence the willingness of hotel guests to reduce disposable items by passing on descriptive or directive norms. This study examines the intermediary role of personal norms (activating the model through norms) and the regulatory effect of Perceived anonymity, aiming to optimize human–computer collaboration, so as to promote sustainable consumption. Design/methodology/approach Two experimental studies (n = 240; n = 480) used a 2 × 2 factorial design (norm type: descriptive/injunctive × service agent: human/robot). Study 1 uses analysis of variance and PROCESS software to test the main effect and the intermediary effect. Study 2 is extended to 2 × 2 × 2 design and incorporates Perceived anonymity. The pretest verified the effectiveness of the experimental operation through the Likert scales. The data analysis adopts the adjustment intermediary model. Findings Robot agents enhance the intention to reduce through descriptive norms, while humans perform better in directive norms. Personal norms play an intermediary role in these interactions. Low anonymity amplifies the advantages of human directive norms, while high anonymity offsets the impact of consistency between the agent and the norm, thus reducing the efficiency of the robot. Originality/value This study took the lead in introducing the concept of agent-standard consistency effect in the hotel service industry. By introducing the concept of anonymity, it expanded the focus theory and normative activation model of normative behavior. It provides a practical strategy for matching service personnel with standard types and promotes sustainable consumption by realizing situation-sensitive human–computer task allocation in the digital service ecosystem.