The contextual integrity model is a widely accepted way of analyzing the plurality of norms that are colloquially called "privacy norms". Contextual integrity systematically describes such norms by distinguishing the type of data concerned, the three social agents involved (subject, sender, and recipient) and the transmission principle governing the transfer of information. It allows analyzing privacy norms in terms of their impact on the interaction of those agents with one another. This paper places contextual integrity in a strict game theoretic framework. When such description is possible it has three key advantages: Firstly, it allows indisputable utilitarian justification of some privacy norms. Secondly, it better relates privacy to topics which are well understood by stakeholders whose education is predominantly quantitative, such as engineers and economists. Thirdly, it is an absolute necessity when describing ethical constraints to machines such as AI agents. In addition to describing games which capture paradigmatic informational norms, the paper also analyzes cases in which the game, per se, does not encourage normative behavior. The paper discusses two main forms of mechanisms which can be applied to the game in such cases, and shows that they reflect accepted privacy regulation and technologies.

翻译：暂无翻译