Optimism and pessimism in trust

Artificial agents do not exist in the world in solitude. They are required to interact with others, and thus must reason in some fashion about those they interact with. This paper presents a view of trust which artificial agents can use to help in such reasoning processes, providing them with a means to judge possible future behaviour on the basis of past experience. The paper discusses the notion of ‘dispositions’ of trusting behaviour, which we call Optimism, Pessimism and Realism. Each different disposition results in different trust estimations from an agent. We discuss the possible effects of these differences. Finally, we present the concept of memory in trusting agents, and briefly suggest some ways in which memory spans can affect the trusting decisions of such agents, with different dispositions.