Recent research in the field of social robotics has shed light on the considerable role played by biases in the design of social robots. Cues that trigger widespread biased expectations are implemented in the design of social robots to increase their familiarity and boost interaction quality. Ethical discussion has focused on the question concerning the permissibility of leveraging social biases to meet the design goals of social robotics. As a result, integrating ethically problematic social biases in the design of robots–such as, e.g., discriminatory gender stereotypes–has been opposed as morally unacceptable. Building on this debate, the present paper explores a related but different question: would it be permissible to design social robots in ways that intentionally challenge widespread discriminatory social biases, thus fostering moral education? The analysis shows that, while the potential benefits of such a design strategy could be significant, its practical endorsement raises important ethical issues. Hence, caution and further discussion are advised.