The legal aspects of actors in information systems are being investigated by Karnow and others. The question is raised whether electronic agents' digital identities should have their own rights and whether it is an infringement of the rights of the client if an agent acting on his/her behalf is being denied access to a service.
What if there are agents that will only service Swedish users? Is this equal to discrimination of nationality? Or what if a person can't use the global Internet market because he/she once made some crime? Even though the person already has paid his debt, he might find himself forever being refused service on arbitrary grounds. This denial of service could be a serious problem for automatic reputation systems.
How should a disagreement between two electronic actors be resolved? Krogh [4] proposes that agent systems should have normative legal systems that should be used when such matters have to be settled. The question is whether or not these rules have to be coded into the system and globally enforced. It would be most satisfactory if this was not the case, but since the Internet surely will have effects on the world outside of it, it might be necessary to make legal systems in different countries be able to work together.
In short, there is need for a way to prove and motivate denial of service as well as electronic contracts that allow unpartial jurors to decide who is being wronged by whom.
An interesting observation is that buyer anonymity can help preventing some kinds of discrimination. If an actor is anonymous there is no way anyone can know the ethnic group, sex or nationality of the client, and therefore discrimination on these grounds is made impossible.
___________________________________________________
Lars Rasmusson