Implicit biases have been studied by social psychologists for almost three decades, mainly as an individual phenomenon. Recent proposals, however, reframe implicit biases as a collective or structural phenomenon. The “Bias of Crowds” framework is one influential such proposal. Its label carries a clear reference to the idea of the “Wisdom of Crowds”. The connection between these two frameworks has, however, hitherto only been stated at the level of superficial analogy, rather than in-depth analysis. Thus, the details of any proposed similarity, and the specific implications for the very idea of a Bias of Crowds, remain as of yet unclear. This paper provides a detailed assessment of the idea of a Bias of Crowds through the Wisdom of Crowds framework. It shows that there are several interpretations of the former expression, which go well beyond what is suggested by the vague analogy. The results are not just conceptual, however, but moreover have explanatory power: the analysis uncovers how collective decision-making can amplify bias. This provides a novel explanation of the observed correlations between average collective bias scores and large-scale structural inequalities, which inspired the Bias of Crowds framework in the first place. An important byproduct of the analysis is the insight that, on an overall picture, amplified collective bias may fly under the radar of amplified collective competence, and thus be nay invisible. Clarifying the idea of a Bias of Crowds in this way answers the crucial question how extremely small-scale biases can have extremely large-scale social effects.