In response to the limitations of comtemporary database management systems in addressing the requirements of many potential application environments, and in view of the characteristics of emerging interconnected systems, we examine research directions involving adding more 'intelligence' to database systems. Three major thrusts in the intelligent database systems area are discussed. The first involves increasing the modeling power to represent an application environment. The second emphasis centers on enhancing the functionality of such a system to include rule-based inference capabilities. The third direction focuses on supporting the interoperability of multiple, autonomous, heterogeneous database systems. We provide here a view of the first two thrusts, and examine the third via a description of the Remote-Exchange research effort currently underway at USC.