This paper presents the research for applying the AI blackboard paradigm to the problem of realistically emulating multiple battlefield entities within a training simulator. One challenge of these computer generated forces (CGF) is to emulate human behavior so human controlled and computer controlled entities are virtually indistinguishable. The blackboard paradigm provides a useful framework for attacking the CGF problem. The results of this research demonstrate the usefulness of a blackboard architecture for the CGF problem. The blackboard paradigm provides a means for integrating subtasks of the system that are implemented using different programming paradigms. Also, the "context" and "event" driven control strategy of the blackboard paradigm provides adaptive behavior for the computer control forces. These characteristics discriminate the blackboard architecture from other programming paradigms for use with the CGF problem. The research reported in this paper was funded by STRICOM under contract N61339-92-C-0032.