An object-oriented architecture in AI is organised as a collection of components (objects/agents) that each have different aspects of knowledge and/or reasoning abilities and interact with one another, typically through messages, in order to solve a problem. The individual comonents need not use oblect-priented programming, so long as they interact in an object-based manner at a high level.
An object-oriented architecture is similar to a blackboard architecture in that both consist of agents cooperating to solve a problem, but in the object-oriented architecture they communicate dircetly with one another, whereas in a blackboard architecture they interact indirectly via the blackboard.
Used in Chap. 16: pages 243, 244