AbstractSo far, the Ambient Intelligence (AmI) paradigm has been applied to the development of a great variety of real systems. They use advanced technologies such as ubiquitous computing, natural interaction and active spaces, which become part of social environments. In the design of AmI systems, the inherent collaboration among users (with the purpose of achieving common goals) is usually represented and treated in an ad-hoc manner. However, the development of this kind of systems can take advantage of rich design models which embrace concepts in the domain of collaborative systems in order to provide the adequate support for explicit or implicit collaboration. Thereby, relevant requirements to be satisfied, such as an effective coordination of human activities by means of task scheduling, demand to dynamically manage and provide group- and context-awareness information. This paper addresses the integration of both proactive and collaborative aspects into a unique design model for the development of AmI systems; in particular, the proposal has been applied to a learning system. Furthermore, the implementation of this system is based on a blackboard- based architecture, which provides a well-defined high-level interface to the physical layer.