The anthropomorphic appearance of intelligent machines has the potential to enhance user acceptance during human-machine interaction but also entails ethical issues. The matching effect of machine gender appearance is considered to involve social stereotypes and gender biases, representing a form of bias alignment. This study explores human-robot compliance from the perspective of ethical sensitivity and finds that while the matching effect between user gender and machine gender does not directly affect compliance, gender bias still exists in human-robot interaction. The degree of anthropomorphism itself, particularly the user’s willingness to anthropomorphize, has a significant impact on compliance. Therefore, it is necessary to distinguish between the user’s willingness to anthropomorphize and the machine’s anthropomorphic elements in robot design as elicited by ethical sensitivity mechanisms.