Openness and disclosure in multi-agent learner models

Multi-agent systems raise many new and interesting research issues in AI-based educational collaborative systems. Some of those issues relate to the user model. When many agents inhabit a system, they must cooperate in their agent society, interacting with each other. They must interact with the world outside the computer (i.e. with the learners). They must have adequate knowledge of the users they represent. These new issues also manifest themselves in new issues for open and inspectable learner models. This paper presents some of these issues in the context of the I-Help project