Our regulatory institutions, broadly taken, include our moral norms and models and have not fully adapted to significant changes in data mining technology. For example, we suggest that the metaphors — Big Brother and “data mining” itself — commonly used to describe and assess this new technology are deficient, overemphasizing social discipline by the state and the passivity of the so-called data subject. We move from metaphors to a set of models more adequate for building an ethics of data mining, using a framework of informal game theory. We sketch three models of interaction: pure conflict, pure coordination, and a mixed motive cooperation game, with special application to security, health, and commerce, respectively. We recommend these three models as heuristics within a simple account of an ethics of data mining regulated by informed consent.