During the last decade, a growing number of research projects both in academia and industry embarked on the development of life-like agents as a new metaphor for highly personalized human-machine communication. Due to advances in computer graphics, the realization of visually appealing agents, either based on real video, cartoon-style drawings or even model-based 3D graphics, has become manageable. The visual appearance of a life-like interface agent is, however, only one side of the coin. To be useful, however, agents have to behave in a reasonable and natural manner.
             An interface agent has been defined as "a computer program to provide assistance to a user dealing with a particular computer application" (Meas, P. et al, 1993). Interface agents can be used to perform a variety of tasks as they communicate with the system user or other agents these tasks may include information filtering and retrieving, scheduling meetings, mail management, device control, etc. More advanced interface agents are termed 'Intelligent' if they learn by "watching over the shoulder" of the user and detect patterns and regularities in the user's behaviour. Intelligence introduces a stronger notion of agency. Intelligence gives an interface agent the ability to learn or understand from experience and make decisions in new situations without interacting with the user.
             Four key characteristics of an agent were defined by Wooldridge and Jennings: 1. autonomy towards the user and the environment, 2. social ability to communicate with other agents, 3, reactivity towards the changes of its environment, 4, proactive behaviour by the way of goal-oriented initiatives. These were expanded, by Wolfgang Wahlstar, as concepts for Agent-based user interfaces, as shown in Figure 1 below:
             Figure 1: Interface Agents characteristics.
             Each of the four arms illustrates perceived required characteristics. To be seen as active, the agents should appear as life...