Diversity structures (e.g. diversity trainings) have been implemented in companies to bring awareness to discrimination and promote equality (Shen, Chanda, D’Netto, & Monga, 2009). However, previous research has shown that diversity structures can act as legitimizing cues and can threaten the authenticity of claims of discrimination made by low-status groups (Dover, Major, & Kaiser, 2012). Ironically, high-status group are typically threatened by pro-diversity organizational messages (Dover, Major & Kaiser, 2015). Could diversity structures meant to help minorities make it more likely that high-status groups are more likely to believe bias claims from an in-group member? In Study 1, White participants were randomly assigned to read a description of a company that had a diversity training versus regular employee training, then read a claim of discrimination made by a White employee. They also completed a measure of White group identification. There was a significant interaction (Condition X GID), which suggested that when participants in the control condition identified more with their in-group, the more believable they found the claim of discrimination to be. The unexpected but interesting result could be due to the fact participants in the diversity condition withdrew their support of the White claimant because they did not want to appear racist or felt that the White claimant was acting as a bad in-group member. A replication of the Study 1 with the additional measures, group level social cost and individual level social cost was conducted, and the results did not replicate. However, exploratory mediation analyses revealed group level social cost served as a significant mediator for the relationship between GID and claim believability, perceptions of diversity structures and perceptions of policy changes while individual social cost did not. Implications for diversity structures, GID and future research directions are discussed.