Information exchange is often regulated by law, such as data protection
and privacy law. This holds in various domains, such as crime
investigation, health care and child care. Since data are increasingly
exchanged online (think e.g. of electronic patient records), it becomes
increasingly important to study the automated enforcement of the relevant
regulations. A key issue in such studies is to reconcile the goals to
promote optimal information exchange and to respect the relevant law.

In this talk this problem is tackled in the form of a multi-agent
architecture for regulated information exchange. The architecture is
applied to the exchange of crime investigation data between police forces
in the Netherlands.  It combines a rule-based argumentation logic with a
BDI-based agent model and a protocol for two kinds of inter-agent
communication. Interactions between police officers about information
exchange are modelled as negotiation dialogues with embedded persuasion
dialogues. An architecture is then proposed consisting of two agents, a
requesting agent and a responding agent, and a communication language and
protocol is proposed with which these agents can interact to promote
optimal information exchange while respecting the law. Finally, dialogue
policies are defined for the individual agents, specifying their behaviour
within a negotiation and persuasion dialogue. The policies are designed in
such a way that the overall goals to optimize information exchange within
the bounds of the law are realised.

The architecture is an example of how techniques from the areas of
defeasible argumentation and multi-agent systems can be combined and
applied in a realistic setting.