Affiliations: Supélec, Campus de Rennes, CS 47601, Avenue de la Boulaie, 35576 Cesson-Sévigné Cedex, France | CNRS - Laboratoire d'Informatique de Grenoble, Maison Jean Kuntzmann, 110 avenue de la Chimie, Campus universitaire, 38400 Saint-Martin-d'Hères, France, E-mail: [email protected]
Abstract: The aim of this study is to provide artificial agents with logical tools to reason specifically on privacy-related regulations, in order to comply with them. In order to express these regulations, the Deontic Logic for Privacy (DLP) is proposed. DLP is a deontic and temporal logic based on predicates dealing with personal data management. Illustrated by an example, it becomes obvious that specific operators are needed to express mixed deontic and temporal notions such as obligations with deadlines and maintained interdictions. A set of eight specific requirements is defined for such operators. The existing proposals in the field are then evaluated with respect to these criteria, and found insufficient. Two new operators are tailored to fit both the DLP formalism and the eight requirements. It is then shown how such operators can be used to translate typical privacy regulations in logical formulae via the DLP language.