<p>Over the last decade, philosophers and the public have raised concerns about the privacy implications of big data analytics. On one hand, contemporary machine-learning and data-mining techniques can be used to make inferences about sensitive information that seem, intuitively, to violate subjects’ privacy. On the other hand, it is unclear how such inferences can violate privacy, so long as they are made using only data that was legitimately acquired. According to most privacy theorists, the moral right to privacy protects against certain means of information acquisition, and if any means of acquiring information respects privacy, then it seems to be inference. In this paper, I aim to reconcile the acquisition thesis—that is, the view that the right to privacy concerns the means by which information is acquired—with the intuition that inferences can violate privacy, even when made on the basis of legitimately acquired data. In order to do so, I argue that the right to privacy can be violated when individuals are subjected to excessive attention. Even information-gathering practices that normally respect privacy can be made to violate it when performed in a way that directs excessive attention at an agent. In short, the right to privacy is partly a right to obscurity, and this right can be directly threatened by the extensive computational resources that big data makes it possible to marshal.</p>