PurposeService robots are gradually becoming more anthropomorphic and intelligent. This research aims to investigate how anthropomorphic service robots with different levels of intelligence affect their human counterparts.Design/methodology/approachTwo between-subject experimental studies were used to test whether different levels of service robot anthropomorphism with different levels of intelligence influence employees' morale and resistance to service robots.FindingsStudy 1 shows that the effect of service robot anthropomorphism (low vs. high) on employees' resistance and morale is mediated by perceived job-security threat. Study 2 validates this mediating effect and shows that it is moderated by the type of AI (mechanical vs. analytical). Specifically, when exposed to mechanical AI-powered service robots, employees exhibit a higher perceived job-security threat toward robots with a high (vs. low) degree of anthropomorphism. This moderating effect is not observed when employees are exposed to analytical AI-powered service robots. This moderated mediation effect is also found for the signing of a petition as the behavioral outcome.Practical implicationsService firms considering the adoption of mechanical AI-powered service robots should choose a low (vs. high) anthropomorphic robot to reduce the sense of job-security threat felt by human employees, which subsequently increases their acceptance. However, if analytical AI-powered service robots with are to replace their human employees, the degree of anthropomorphism becomes irrelevant.Originality/valueThis is the first empirical study to explore how anthropomorphic service robots can influence human employees' evaluations and behaviors.