Politeness is commonly defined as a system of interpersonal relations designed to facilitate interaction by minimizing the potential for conflict and confrontation inherent in all human interchange (cf. Lakoff 1973). Building on this, Brown and Levinson’s (1987) classic theory of politeness centers on the concept of protecting one’s “face” or image in a public domain. This study explores whether humans extend the use of polite interaction strategies to situations in which they collaborate with non-anthropomorphic robots. Using a Wizard-of-Oz experimental setup, native German speakers were instructed to assemble an IKEA shelf with either a robotic arm supposedly equipped with a conversational agent or a human partner. Audio recordings of the interactions were transcribed and annotated to analyse the use of linguistic mitigators and to compare politeness strategies across both conditions. Results showed that participants used more impositives and conventionally indirect strategies when interacting with the robot, whereas interactions with humans featured more information requests and nuanced mitigation. This research contributes to the growing field of human-robot interaction by providing empirical evidence on how social norms such as politeness are maintained or altered when interacting with artificial agents. The findings have implications for the design of socially aware, non-anthropomorphic manufacturing robots that can engage in natural and culturally appropriate communication. Full text …