It's not stealing if you need it: a panel on the ethics of performing research using public data of illicit origin

In a world where sensitive data can be published to a worldwide audience with the press of a button, researchers are increasingly making use of datasets that were publicized under questionable circumstances. In many cases, such research would otherwise not be possible. For instance, Weir et al. examined over thirty million user-generated passwords in order to observe the effects of entropy on password cracking [10].