Anthropomorphic robot features are intended to trigger the activation of social scripts from human-human interaction, thereby offering an intuitive approach to interact with robots. Whereas this seems to be a valid design option for the social domain leading to an increased acceptance of robots, trust and willingness to interact, other domains of human-robot interaction (HRI) do not provide such a clear picture. In my talk I will address the impact of anthropomorphic robot design on trust and behavior in work-related settings and show the drawbacks of anthropomorphism when not considering the context of HRI and the task relevance of anthropomorphism.