Protecting e-commerce agents from defamation

摘要

A new theoretical model for the control of defamation in virtual environments is proposed. The goal of the proposed anti-defamation model is to develop effective mechanisms to protect e-commerce agents from defamation. The proposed model differs from a more general deception handling model in that only data relating directly to the calculation of reputations are evaluated for deception. Defamation is defined as the unfair damage of an individual’s (or group’s) reputation in the view of the right-thinking majority. To be more specific in order to be classified as defamation the damage must be caused by false statements made to a third party. In order for an individual’s reputation to be damaged their image must be lowered in the mind of another or several others. By showing the potential for damage in systems with reputation-type mechanisms this work stresses the need for anti-defamation measures. These measures are necessary to protect the innocent against malicious attacks on reputations. An e-commerce simulation based on this model is described and results are discussed. Results from the experimental simulations support the validity of the proposed anti-defamation model as they show the anti-defamation model is successful in protecting the simulated society from the overall damage caused by defamation. A concerning outcome, however, is that while all agents are better off with the proposed anti-defamation model the model is unsuccessful in targeting and punishing specific dishonest agents. Although this work is directed toward the specific area of defamation we believe it may be of particular interest to anyone wishing to provide protection against all types of deception.