This paper presents the results of research whose aim was to determine the type and amount of personal information individuals might disclose to robots designed with different visual appearances. The set of images viewed by participants consisted of two humanoid appearing robots and a female android. Further, a human image was used as a control for comparison purposes. For an individual to decide to disclose personal and potentially embarrassing information to a robot serving as a counselor, they must trust that the robot will safeguard their disclosures and be an empathetic listener. In this research 110 participants viewed four images and completed an online survey accessing their attitudes and decision on whether to self-disclose personal information to a robot counselor. Compared to the robot images, the results indicated a strong preference to disclose personal information to a human counselor regardless of the type of information. However, given the type of self-disclosure, the data also showed that participants would, to some extent, disclose to a friendly appearing robot and female android, and more so than to a robot judged to lack affect.